Bulletin of the American Physical Society
63rd Annual Meeting of the APS Division of Plasma Physics
Volume 66, Number 13
Monday–Friday, November 8–12, 2021; Pittsburgh, PA
Session PP11: Poster Session VI:
Poster

Hide Abstracts 
Room: Hall A 

PP11.00001: LOWTEMPERATURE PLASMA


PP11.00002: Investigation of He spectra using a collisionalradiative model for plasma diagnostics of dielectric barrier discharges Shota Abe, Surabhi Jaiswal, James B Hall, Bruce E Koel, Ahmed Diallo We have performed optical emission spectroscopy to measure He spectra from pure He, or N_{2} and/or H_{2} mixed He dielectric barrier discharge plasmas for near atmospheric pressures (APs) up to 200 Torr. Two reactor configurations, coaxial cylinder and parallel plates, were examined for an applied AC voltage of 15 kV at 20 kHz. The obtained He line intensity ratios were investigated using a He collisionalradiative (CR) model to determine the plasma parameters T_{e} and n_{e}, which are critical parameters to understand plasmaassisted catalysis for ammonia synthesis. The He CR model, originally developed by Goto [1], was integrated with AP reactions such as heavyparticle collisions between He atoms and molecules [2]. Our model analysis shows that He excitedstate populations for AP plasma are strongly affected by the population density of metastables He(2^{1}S) and He(2^{3}S), which were treated as free parameters determined together with T_{e} and n_{e}. We utilized a zerodimensional kinetic model developed for N_{2}H_{2}He AP plasmas to estimate these metastable populations. [1] M. Goto, J. Quant. Spectrosc. Radiat. Transfer 76, 331 (2003). [2] W. Lee et al., Phys. Plasmas 27, 073502 (2020). 

PP11.00003: Metastable Oxygen Formation in an Atmospheric Pressure Plasma Jet Evan M Aguirre, Surabhi Jaiswal Presented here are detailed measurements of an argon atmospheric pressure plasma jet producing metastable oxygen through the (O(^{1}S)O(^{1}D)) transition. The emission at 557 nm, which is also referred to as the “auroral green” line, is observed regardless of electrode design, and various operating parameters such as applied voltage, gas flow and electrode gap. The plasma jet is characterized in detail using optical emission spectroscopy and analysis of electrical diagnostics. The intensity of oxygen and other reactive species are monitored with a high resolution spectrometer and CCD camera. The temporal behavior of the jet is monitored with a fast imaging CMOS camera. We also present analysis of the mean electron energy, electron density, and gas temperature. The stable behavior of the jet provides a large surface area for plasma treatment. Lastly, we present evidence of the plasma jet’s ability for industrial applications such as wastewater treatment. 

PP11.00004: XPinch source characterization for the study of material behavior under shock loading towards Xray diffraction Camille Chauvin, David Palma de Barros The use of XPinch generators constitutes an alternative way to implement fast intense Xray emission to study short phenomenon towards radiography or Xray diffraction. CEA Gramat has developed a XPinch source coupled to a single stage gas gun which is able to perform Xray diffraction under shock loading. The first experiments showed promising results, however, the source intensity profile has shown significant differences. A better understanding of the source emission over 10 keV in terms of size and spectral emission is needed to optimize the diffraction parameters. These characterizations were done using loads made of gold wires of 25 µm diameter previously used for diffraction. A crystal spectrometer and pointprojection imaging were both implemented at each flash and the results compared. Characteristic time integrated spectra were obtained with a spatial integration of the crystal spectrometer image. Pointprojection imaging showed that up to three different vertically shifted sources could be involved in the emission. Classical geometrical relations are not adapted in this case and a specific tool was designed to simulate the spatial profile overlap and try to reproduce experimental measurements. The first simulations showed that the different source diameter spread from several hundred microns to over a millimeter. The intensity profile seems to influence weakly the measurements. This first study showed encouraging results and further experiments are planned to validate the tools. A database could be then built over various material wires and diameter. 

PP11.00005: Effect of noble gas addition on plasmaassisted ammonia synthesis Zhe Chen, Surabhi Jaiswal, Bruce E Koel Ammonia synthesis assisted by nonthermal plasmas has gained increasing interest because nonthermal plasmas allow ammonia synthesis to occur at atmospheric pressure and temperatures lower than what is required by the HaberBosch process [1]. We report here the effect of noble gas addition on plasmaassisted ammonia synthesis. We performed kinetics experiments and optical emission spectroscopy (OES) measurements using a coaxial dielectric barrier discharge (DBD) reactor at several applied voltages with different percentages of Ar and He added to the N_{2}H_{2} reaction mixture. Our kinetics experiments showed that the NH_{3} mole fraction produced reached a maximum for 10% Ar or He, and that He addition gave a higher NH_{3} mole fraction than Ar. OES data showed the existence of reactive gas phase species (N_{2}^{+}, N, H_{α}, H_{β}, and N_{2}^{*}) in both cases, with a larger amount of H observed for He addition. Atomic hydrogen lines H_{α} and H_{β }were also utilized to estimate the plasma parameters. [1] J. G. Chen et al., Science 360, eaar6611 (2018). 

PP11.00006: Characterization of an atmospheric pressure carbon arc plasma Nirbhav S Chopra, Yevgeny Raitses, Shurik Yatom, Jorge Munoz Burgos Atmospheric pressure anodic carbon arc discharges are a promising method for lowcost, highvolume synthesis of nanomaterials. During arc operation, carbon material is introduced into the arc by the ablation of the graphite anode [1]. The anode ablation depends on the power balance at the anode, which is influenced by whether the anode sheath is electronrepelling (negative anode sheath) or electronattracting (positive anode sheath) [1–4]. Anodic carbon arcs exhibit a transition between low and high ablation modes; at larger arc currents the ablation rate of the anode grows nonlinearly [1,3]. We show the existence of a positive anode sheath in both low and high ablation modes. The electron temperature and density are determined by optical emission spectroscopy and corroborated by a Langmuir probe measurement. The plasma potential is determined with a floating probe. The floating probe potential is related to the plasma potential by assuming ions diffuse through neutrals in the probe presheath. Effects of the positive anode sheath on anode ablation rate are discussed. We also discuss a plausible explanation for the discrepancy in experimentally determined discharge voltage and discharge voltage calculated by recent models of the arc [4]. 

PP11.00007: Measuring azimuthal magnetic field magnitudes in a plasma gun generated single flux rope with laser induced fluorescence Tyler J Gilbert, Katey J Stevenson, Earl Scime It has been shown that laser induced fluorescence measurements of Zeeman split spectra offer a method to nonperturbatively measure magnetic fields in laboratory plasmas.^{1,2} Laser induced fluorescence is a nonperturbative laser spectroscopic technique that uses the Doppler motion of a species and a narrow linewidth laser to measure the velocity distribution function of the particles. Prior measurements demonstrated a magnetic field resolution of 10 G is achievable.^{2} For the results presented here, a Toptica diode laser is freespace injected parallel to the background guiding magnetic field into the PHAse Space MApping experiment (PHASMA), a helicon plasma source equipped with two plasma guns capable of generating 10 ms long flux ropes. Zeeman split Ar I σpeaks are measured in a single flux rope in the presence of a background helicon plasma. Here we present measurements of the azimuthal magnetic field created from a single flux rope. These initial measurements show the viability of using this technique for future magnetic imaging of laboratory magnetic reconnection events arising from the merger of two flux ropes. 

PP11.00008: Electroncollision Helium Ion State Enhancement from Electron Beam for He II Twophoton Absorption Laser Induced Fluorescence Thomas E Steinberger, Jacob W McLaughlin, David D Caron, Earl Scime Speciesspecific investigation of particles using laser spectroscopic techniques rely on sufficient state population and transition wavelengths that are suitable for available lasers, vacuum components, and detection electronics. Commonly, laserbased fluorescence diagnostics target metastable or ground states to ensure ample initial state density. However, the transitions are often very energetic. Twophoton absorption laser induced fluorescence (TALIF) is ideal for probing energetic transitions from metastable and ground states in many species, but, some species, such as helium, are difficult to investigate with this technique since ground state He ion transitions are too energetic to be probed with TALIF laser wavelengths (~30 nm) and lowlying energy state metastables are rarely naturally populated in cold plasmas. Here we populate the 2s metastable state of singly ionized helium directly by injecting an energetic electron beam into a background helium plasma. The 2s state is then pumped to the 6s state by two 205 nm photons and fluorescence intensity from the 6s to the 4p decay is collected at ~656 nm. We present, for the first time, He II velocity distribution functions (IVDF) measured with TALIF in an electron beamenergized helicon plasma. Helium ion temperature and bulk flow are determined from the IVDFs. 

PP11.00009: Intermodulated Laser Induced Fluorescence Measurements of Singly Ionized Atomic Iodine Matthew J Lazo, Thomas E Steinberger, Timothy N Good, Earl Scime Electrostatic propulsion systems widely use xenon as a propellant, however, iodine is one candidate under consideration to replace xenon due to its attractive qualities as a propellant. A key tool missing in the development of iodinefueled propulsion systems is a spatially resolved diagnostic technique to analyze ion flow rates and temperature of the thruster plasma, in order to characterize the performance of such systems without perturbing the plasma. Previous work investigated the lineshape of singlyionized atomic iodine (I II) with laser nduced fluorescence (LIF), but could not resolve the hyperfine structure of the ^{5}D^{0}_{4} and ^{5}P_{3} states [Steinberger and Scime, Journal of Propulsion and Power, 34, 2018]. Here, an ntermodulated LIF technique is used to measure a Dopplerreduced lineshape of the same I II transition. A Monte Carlo fitting algorithm is used to fit the transition lineshape, where the magnetic dipole and electric quadrupole coupling coefficients are left as free parameters with constraints from theory. We report most probable values of the hyperfine coupling coefficients for these I II states. 

PP11.00010: Plasma Diagnostics Measurements on the Helicon Plasma Experiment (HPX) Royce W James, Richard W Freeman, Lorraine A Allen The small Helicon Plasma Experiment (HPX) at the Coast Guard Academy Plasma Lab (CGAPL), continues to progress toward utilizing the reputed high densities (10^{13} cm^{3} and higher) at low pressure (.01 T) of helicons, for eventual high temperature and density diagnostic development in future laboratory investigations. HPX has installed an Impedans Langmuir probe and constructed an RFshielded triple probe experimental diagnostic to compare the plasma's density, temperature, and behavior during experiments. Our 2.5 J YAG laser Thomson Scattering system operates at its first and second harmonic, 532 and 1064 nm respectively. It utilizes a highperformance volumephaseholographic (VPH) grating spectrometer and a chargecoupled device (CCD) camera with a range of 3801090 nm with a resolution of 1024x1024 for second harmonic (532 nm) photon emissions. At 1064 nm, a new polychromator has been procured from General Atomics optimized for TS measurements of 5 eV < T_{e} < 2000 eV over a 109degree scattering angle. Preliminary observations from the Thomson Scattering, particle, and electromagnetic scattering diagnostics will be reported. 
Not Participating 
PP11.00011: Measurement of electron temperature in an E x B Penning source by optical emission spectroscopy V S Santosh K K Kondeti, Shurik Yatom, Yevgeny Raitses E x B discharges are known to exhibit strong low frequency oscillations such as breathing oscillations and rotating spokes.[i] There is a strong interest and need for the characterization of these oscillations and their effect on the plasma using minimuminvasive and noninvasive diagnostics. In this work, we performed high resolution emission spectroscopy to obtain the spatially resolved electron temperature in the plasma generated in an E x B Penning source. We used the line ratio of the emission from the Ar (3p_{5} > 1s_{4}) and Ar (3p_{9} > 1s_{5}) lines at 419.83 nm and 420.07 nm respectively to obtain the electron temperature. This ratio depends on the electron temperature and the ratio of the Ar (1s_{5}) metastable density and the ground state Ar density.[ii] The Ar metastable density was obtained using absorption spectroscopy and the ground state Ar density was obtained from the pressure of the chamber.[iii] The obtained electron temperature was validated using Langmuir probe measurements. The obtained trends of the electron temperature measured with OES are consistent with that measured with the probes. [i] E. Rodriguez, V. Skoutnev, Y. Raitses, A. Powis, I. Kaganovich, and A. Smolyakov, Phys. Plasmas 26, 053503 (2019) [ii] J. B. Boffard, R. O. Jung, C. C. Lin, L. E. Aneskavich, and A. E. Wendt, J. Phys. D: Appl. Phys., 45(4), 045201 (2012). [iii] J. B. Boffard, R. O. Jung, C. C. Lin, and A. E. Wendt, Plasma Sources Sci. and Technol., 18(3), 035017 (2009). 

PP11.00012: Validation of Collisional Radiative Models for Low Temperature Xenon Plasmas Mary Konopliv, Anirudh Thuppul, Richard E Wirz Collisional radiative (CR) models for neutral xenon (Xe I) and singlycharged xenon ion (Xe II) have been recently developed by Chaplin, et al. This poster will cover recent work to experimentally validate the Xe II CR model for future nonintrusive measurements of plasmas with densities of 10^{18} m^{3} and temperatures of 10 eV. Additionally, efforts to extend this work to timeresolved OES will be discussed. Discrepancies between the existing Xe I CR model developed by Karabadzhak et al. and Chaplin et al. are explored, and explanations for these discrepancies are provided to justify the structure and theoretical database used in both the new Xe I and Xe II CR models. An experiment to validate the Xe II CR model was carried out in the anode region of the UCLA Plasma interactions facility by comparing optical emission spectroscopy (OES) with the Xe II CR model with Langmuir probe measurements. The spatially resolved electron temperature and density measurements show reasonable agreement between the two diagnostics. The 441 nm to 605 nm emission line ratio appears to have the most promise for nonintrusive electron temperature measurements with the Xe II CR model. Also, the bandwidth limitations of the Xe II CR model have been explored, and the 441 nm to 605 nm line ratio may be useful for timeresolved optical measurements as well. 

PP11.00013: Development of probe diagnostics for EM wave measurements in the ALEXIS and MDPX devices Jared C Powell, Edward E Thomas, Saikat Chakraborty Thakur


PP11.00014: Comparison of Electron Density Measurements Using Hairpin and Langmuir Probes in a Magnetized Plasma Yhoshua Wug, Patrick Pribyl, Troy A Carter Hairpin probes are small microwave resonant structures that are commonly used to measure electron density in plasmas. The Ushaped resonant structure undergoes a shift in frequency when immersed in plasma that is related to the plasma frequency, from which the electron density is obtained. Hairpin probes constructed so far can measure plasma density up to ~10^{12} cm^{3 [1]}, which corresponds to a plasma frequency of about 9 GHz. In magnetized plasmas, swept Langmuir probes are known to perturb the plasma being measured, depleting flux tubes they intercept. This leads to erroneous conclusions such as studying the depletion rather than the ambient plasma. Uncoupled hairpin probes do not draw any current, hence flux tube depletion is not a sizable source of error in electron density measurement. We describe simultaneous measurements using a planar Langmuir probe together with a hairpin probe to measure electron density in the Large Plasma Device (LAPD) at the Basic Plasma Science Facility (BaPSF). Detailed electron density measurement comparison between the two methods will be presented, including using the hairpin probe to document plasma perturbations caused by the Langmuir probe i.e., induced downstream flux tube depletion. We will also discuss the prospect of obtaining higher time resolution temperature data using a dual probe. 

PP11.00015: Characterization of plasma in RF jet interacting with water: Thomson scattering versus Spectral line broadening Shurik Yatom, Trey Oldham, Elijah J Thimsen, Yevgeny Raitses This work reports on comparison of plasma characteristics measured by means of Thomson scattering and spectral line analysis. Both measurements were performed on RF powered plasma jet, running at 20% duty cycle, with Ar as a carrier gas. The plasma was impinging on a water surface located 1 cm away from the powered needle electrode. To support the plasma analysis, we also have performed gas temperature measurement by means of laser Rayleigh scattering and detailed timeresolved imaging of the plasma filament. The difference between the plasma parameters measured by the two methods is discussed and analyzed. The main observed difference is in the reignition of the jet following the voltage switchoff. 

PP11.00016: Development of a Parallel MultiDimensional GridBased Vlasov Equation Solver Chen Cui, Joseph Wang We present a parallel, multidimensional gridbased Vlasov solver(Vlasolver) for collisionless plasma simulations. The VlasovPoisson system is solved using the thirdorder Positive Flux Conservation(PFC) scheme(Filbet, 2001; Umeda, 2008). The parallelization of the code is implemented using domain decomposition and MPI. We present two application studies to demonstrate the capabilities of Vlasolver. The first one is a reevaluation of onedimensional expansion of collisionless plasma expansion into vacuum. The second one is a twodimensional simulation of plasma plume emitted from plasma thrusters. It is shown that the gridbased method eliminates the inherent statistical noise in particlebased methods and allows us to extend the solution beyond the selfsimilar expansion region and resolve the effects from electron time scale wave perturbations. We find that the gridbased Vlasov method is advantageous over particlebased methods for simulations of extremely nonhomogeneous plasmas due to its capability to eliminate statistical noise in higher order velocity moments in the lowdensity region. 

PP11.00017: 0D Kinetic Modeling of Noble Gas Mixed N_{2}H_{2} Plasmas for Ammonia Synthesis James B Hall, Shota Abe, Zhe Chen, Zihan Lin, Surabhi Jaiswal, Ahmed Diallo, Bruce E Koel J.B. HALL, Z. LIN, S. ABE, Z. CHEN, S. JAISWAL, A. DIALLO (PPPL), B.E. KOEL Princeton U.  

PP11.00018: Investigation of sheath electron temperature in RF capacitivelycoupled plasmas using PIC simulations. Michael Q May, Alexander V Khrabrov, Dmytro Sydorenko, Igor Kaganovich Radiofrequency capacitivecoupled (RFCC) discharges have increasing applications in semiconductor manufacturing, plasma propulsion, and nanomaterial fabrication, but many aspects of their operation remain poorly understood. Popular low pressure (< .1 Torr) models for RFCC plasmas assume homogeneous electron temperature (ET) [1], even in inhomogeneous models [2], and a recent model of intermediatepressure (.2 – 6 Torr) RFCC plasmas assumes strictly lower sheathedge ET than bulk ET [3]. Experimental evidence has shown, however, that at high pressure (200 Torr) ET peaks at the sheath edge [4]. Using the 1D particleincell (PIC) code EDIPIC, and the recently developed 2D PIC code EDIPIC2D, sheath electron temperature in low and intermediatehigh pressure regimes is characterized and compared with these models and previous experimental results. 

PP11.00019: Sensitivity analysis of the particle in cell (PIC) code hPIC for near surface plasma conditions. Mohammad Mustafa Sensitivity Analysis (SA) and Uncertainty Quantification (UQ) are used to assess and improve physics models. After assessing the uncertainties produced by input parameters, the next logical step is to carry out sensitivity analysis to ensure both robustness of the code over a wide range of plasma conditions (input parameters) and help understanding the underlying physics. In this work, several methods of sensitivity analysis were used to quantify the sensitivity of the Ion Energy Angle Distributions (IEADs) at the plasma sheath edge. The sensitivity analysis methods utilized include One At a Time (OAT) for local sensitivity analysis and Morris screening and variance decomposition (Sobol indices) for global sensitivity analysis. The particle in cell (PIC) code utilized for the analysis is hPIC. Due to the high computational cost of PIC simulations a surrogate model has been developed and used to generate samples for the sensitivity analysis. The global sensitivity analysis showed that the electron temperature has the most impact on the IEADs followed by the ion temperature and magnetic field inclination angle. 

PP11.00020: Diagnostic of NeAr mixture plasma using finestructure resolved collisional radiative models Rajesh Srivastava, Shubham S Baghel, Reetesh K Gangwar A collisional radiative model has been developed for the diagnostics of NeAr mixture plasma. The 40 lowest enegy levels of Neon corresponding 2p^{5}3s, 2p^{5}3p, 2p^{5}3d, 2p^{5}4s and 2p^{5}4p along with ground state (2p^{6}) and Ion state (2p^{5}) have been taken [1]. Processes viz. electron impact excitationdeexcitation, radiative decay, ionization, three body recombination, self absorption correction, metastable diffusion have been considered. The quenching of Neon 1s level by Ar atoms through associative and penning ionization is efficiently included in the model. The model is coupled with emission and absorption measurements of Boffard et al [2] and plasma parameters viz. electron temperature and electron density are extracted for 1, 6, 14, 25 and 40% Ar into NeAr mixture plasma. These are compared with Languimir probe measurements[2]. The 1s level populations and 2p1s line intensities obtained from our model are compared with the spectroscopic measurements [2]. The self absorption correction (escape) factors and electron impact excitation rates for all the NeAr mixture concentrations are also presented. 

PP11.00021: Analytical and numerical characterization of a one and twodimensional plasma photonic crystal with smooth variations in density W. R Thomas, Uri Shumlak Plasma photonic crystals (PPCs) have the potential to significantly expand the capabilities of current microwave and millimeter wave filtering and switching technologies by providing high speed (μs) control of energy bandgap/pass characteristics in the GHz through low THz range. Furthermore, plasmabased devices can be used in higher power applications than their solidstate counterparts without experiencing significant changes in function or incurring damage. Plasmas with periodic variations in density result naturally from instabilities or self organization, or can be created intentionally. In either case, due to their gaseous nature plasmas cannot support discontinuous density profiles and necessarily have finite density gradients. In this work a dispersion relation is derived for one and twodimensional cold plasma photonic crystals with an arbitrary density profile, and is validated against a discontinuous Galerkin (DG) finite element model (FEM) solution of the same problem. The dispersion relation is then used to quantify the effect of various density profiles (sinusoidal, elliptic sinusoidal, piecewise constant, and piecewise linear) on dispersion characteristics. 

PP11.00022: Characterization of mesoporous silica packed bed reactors for plasma synthesis of ammonia Sophia Gershman, Henry Fetsch, Fnu Gorky, Maria Carreon Plasmacatalytic ammonia production at atmospheric pressure has the advantage of small scalable design, easy intermittent operation, and use of renewable energy, but to meet the ammonia production it needs intelligent design of catalytic materials with different properties than traditional catalysts. One of our main hypotheses is that since plasma aides in the dissociation of N_{2} trough vibrational excitations and produces H* radicals, the catalysts need to adsorb H* to allow it to react with plasma activated nitrogen, and bond to nitrogen weakly. Mesoporous oxides such as silica (SBA15) have these properties and also produce a stable plasma due to their low conductivity, are chemically stable, and have a porous structure that facilitates diffusion and surface reactions. We use optical emission spectroscopy (OES), Fourier transform infrared absorption spectroscopy (FTIRAS), and electrical measurements to characterize two plasmacalytic packed bed dielectric barrier reactors, one packed with SBA15 and another with SBA15 with 10% Ag, operating in a 1:1 N_{2} and H_{2} gas mixture. OES analysis shows similar parameters for two reactors with a 10% higher vibrational excitation temperature and lower rotational temperature in SBA15 reactor as compared to SBA15Ag. At the same applied voltage SBA15Ag produces <20% higher concentration of NH_{3}, but the highest 11,000 ppm of NH_{3} was for SBA15 reactor. Plasma characterization can help choose the catalysts for ammonia production. 

PP11.00023: Effect of surface porosity on plasma assisted ammonia synthesis Surabhi Jaiswal, Zhe Chen, Bruce E Koel, Ahmed Diallo We have investigated the effect of surface porosity of catalytic support particles on plasma assisted ammonia synthesis from nitrogen and hydrogen. The experiments were performed using an AC coaxial dielectric barrier discharge (DBD) plasma reactor at room temperature and near atmospheric pressure (550 torr). Reactor performance was evaluated using support particles of porous silica (SiO_{2}) and nonporous soda lime glass beads of equal diameter. N_{2} conversion, ammonia synthesis rate, and energy yield measured at several applied voltages were found to be higher in the case of porous silica compared to nonporous glass. The effect of these different catalytic supports on the physical properties of the discharge was negligible. High resolution optical emission spectra (OES) were used to explore the evolution of reactive gas phase species N_{2}^{+}, N, H_{α}, H_{β}, and N_{2} in the presence of both support particles. The relative concentration of these reactive gas phase species was higher in the case of the nonporous glass supports regardless of applied voltage, which suggests an increased role for these species in gas phase reactions for this support. However, the higher formation rate of ammonia for porous silica supports indicate the importance of these support surfaces in the plasmacatalytic synthesis of ammonia. 

PP11.00024: Production of Solvated Electrons by Atmospheric Pressure Plasma Jets Adam D Light, Shalese M Lovell, Brian L Henning, Dzafer Camdzic, Quinna Phillips, Benjamin Modlin, Annika Zettl Contamination of ground water with polyfluoroalky substances (PFAS) is increasingly recognized as a major environmental issue. While many advanced oxidation and reduction methods are being explored, lowtemperature plasma technologies offer particularly promising avenues to remediation. Our new lab at Colorado College is being built to study the application of atmospheric pressure plasma jets (APPJs) to PFAS contamination in collaboration with the Fountain Valley Water Project. Recently, it has become clear that free electrons dissolved in water ("solvated electrons") are particularly effective in their ability to breakdown and mineralize PFAS. Atmospheric pressure plasma has been demonstrated to produce solvated electrons in direct discharge at the water/plasma interface in studies by the Rumbach group at Notre Dame, but much remains to be understood. We present progress to date in identifying solvated electrons produced by APPJs, including the design of diagnostics, development of plasma sources, and preliminary measurements. 

PP11.00025: Fabrication of waveguides in flexible glass via femtosecond laser micromachining and visualization of ultrafast dynamics of the laserglass interaction Jack W Agnes, Garima C Nagar, Dennis Dempsey, Nicole A Batista, James S Sutherland, Bonggu Shim We fabricate waveguides in Corning® flexible glass using Femtosecond Laser Micromachining (FLM) and visualize the ultrafast plasma dynamics which lead to waveguide formation via timeresolved interferometry. Due to minimal thermal effects and highlynonlinear optical processes [1], FLM is an ideal tool to fabricate waveguides in glass with high precision and without post processing. We optimize laser fabrication of waveguides by varying scanning speed and pulse energy and, in particular, achieve waveguides with circular crosssections using slit beam shaping [2]. Further optimization requires investigation of the underlying dynamics of how structural changes in glass are made during and after laserglass interactions. Thus, we visualize the creation and recombination of plasma in glass which leads to the formation of waveguides using time resolved interferometry [3]. 

PP11.00026: Acidithiobacillus ferrooxidans: A Natural Substitute for Nitric Acid Leaching of Copper Lynne A Goodwin, Thomas Day, Kumkum Ganguly, Seychelles Voit, Joseph H Dumont, Chris Yeager, Alex Edgar, Steven Young, Alexandria Strickland, Brian Patterson, Tana Morrow Many target components used in plasma physics experiments performed at the Omega Laser Facility (Omega) and the National Ignition Facility (NIF) require the use of nitric acid leaching. These components, specifically gold hohlraums and plastic spools with embedded aluminum layers, are fabricated by machining copper mandrels to the required internal shape and plating or casting the final material onto the mandrel. The final part is then nitric acid leached to remove the copper mandrel. Unfortunately, the optimized leaching process for spools leaves trace amounts of copper. Furthermore, leaching with nitric acid to remove the copper degrades the epoxy and it can begin to attack the embedded aluminum layer, which is critical to the experiment. In this study, a natural bioleaching microorganism, Acidithiobacillus ferrooxidans, was used to augment the current nitric acid leaching process. After the primary leaching process, spools containing small amounts of remaining copper contamination were exposed to a culture of A. ferrooxidans and monitored to test its effectiveness in removing the remaining copper from the spools. This presentation will give an overview of the leaching process and a detailed discussion of the bacteria and bioleaching strategies and challenges. LAUR2126779 

PP11.00027: A generation of hydrogen atoms in DCRF ExB plasma source for advanced material processing applications Yevgeny Raitses, Arthur Dogariu, Fang Zhao, KaiMei C Fu, Christopher G Tully Electron beamgenerated plasmas enable precise control over the flux of species and the ion energy at surfaces making such low temperature plasmas attractive for advanced material processing applications including, but not limited to atomic layer etching and processing of ion energysensitive materials and atomically thin 2D materials [1]. In the ExB Penning plasma system [2], energetic electrons are extracted from the RF plasma cathode to a cylindrical geometry chamber by applying a DC bias voltage between the cathode and the chamber. It is experimentally shown that in operation with argonhydrogen gas mixtures at 110 mtorr, this plasma source can sustain efficient dissociation of hydrogen molecules and ionization of argon and hydrogen atoms. Spatially resolved measurements of the absolute density of hydrogen atoms using twophoton absorption laserinduced fluorescence diagnostic revealed that the density of hydrogen atoms in the chamber substantially increases with the magnetic field. This result and its practical implications for materials processing will be discussed. For example, this source was successfully applied for a high coverage graphene hydrogenation with a minimized irreversible damage to the 2D substrate [3]. [1] S. G. Walton et al., ECS J. Solid State Sci. Technol. 4, N5033 (2015); [2] E. Rodriguez, V. Skoutnev, Y. Raitses, A. Powis, I. Kaganovich, and A. Smolyakov, Physics of Plasmas 26, 053503 (2019); [3] F. Zhao, Y. Raitses, X. Yang, A. Tan, C. G. Tully, Carbon 177, 244 (2021). 

PP11.00028: Spectral characterization of the Hall Effect thruster operating on Kr propellant Oleg V Batishchev, Alexander Hyde, James Szabo Krypton has been viewed by many as a feasible alternative to Xenon gas propellant for EP thrusters. Nowadays, Kr HETs are being used commercially, though they have not been studied as intensively as their Xe analogs. We are reporting our initial results for Busek’s 200W thruster running on Kr propellant, which follow our previous study [1] for Xe. Broad NUVNIR spectra of the hollow cathode and anode discharges are presented first. Next, detailed EUVVIS spectra are discussed. Particular attention is given to strong BIII EUVMUV emission lines and their dependance on the applied power. The highresolution spectroscopic system [2] is used to measure axial velocity and possible azimuthal rotation of plasma exhaust via the Doppler Effect of the prominent ionic emission lines. Experimental data are presented and discussed. 

PP11.00029: Application of permanent magnet assemblies to RF and Arcdriven discharges Alexander Hyde, Chase Leffers, Timofei Kuzmenkov, Oleg V Batishchev We are reporting progress in the development of solenoidlike magnetic field configurations using assemblies of permanent magnet. It was demonstrated that ring assemblies of axially magnetized Nd cylinders create ~0.5T uniform fields in the bores [1]. However, the high field domain it is terminated by magnetic cusps at the exits. We propose a few configurations that can mitigate these singularities to form an open magnetic field topology. The first configuration is an umbrellalike expandable structure of ~30cm long rod magnets, producing B~0.1T [2]. It was used to facilitate continuous rfdriven helicon Ar/N2 discharges [3] and exhaust plasma axially. The second assembly is a reconfigurable chain of 3Dprinted magnet holders that has several degrees of freedom. Six chains are anchored together to form a magnetic mininozzle with B~0.2T at the high end. The enhancing effect of added ferromagnetic wedges is described as well. A short nozzle is used to facilitate pulsed arcdriven discharge and direct the metal plasma flow. 

PP11.00030: A testbed for the Cathodic Arc aluminum plasma thrusters Timofei Kuzmenkov, Asher Solnit, Oleg V Batishchev An increased interest in miniature Cathode Arc Metal Plasma Thrusters was prompted by advances in highpower electronics with application to the CubeSat technology, where compact high efficiency thrusters can significantly extend the mission envelope. Two notable examples are George Washington University’s twoelectrode with applied magnetic field [1] and Alameda Applied Sciences Corporation’s tripleelectrode no field [2] thrusters. Our initial design is like the latter with added ~0.10.3T axial magnetic field, which is created by an external electromagnet. The cylindrical cathode, which serves as the main source of propellant, is made of aluminum due to it representing a majority of the approximately 2kT of space debris in LEO. The design of the driving circuit for the thruster and discharge characteristic data are discussed along with other key elements. Spectroscopic measurements show a high degree of ionization in the exhaust, with singly and doubly ionized species being present. Doppler spectroscopy showed ~17km/s exhaust velocity. Data for different regimes are presented and discussed. 

PP11.00031: Thrust Stand Measurements of an RMFFRC Thruster with Continuous Power System Kenneth E Miller, James R Prager, Alex Henson, Kyle McEleney, Joshua M Woods, Christopher Sercel, Tate Gill, Eric Viges, Benjamin Jorns Eagle Harbor Technologies, Inc. (EHT) developed a continuouslyoperating repetitivelypulsed power system for the University of Michigan (UM) Rotating Magnetic Field (RMF) – FieldReversed Configuration (FRC) Thruster. EHT leveraged a previously designed resonant full bridge in singleburst mode to develop a continuouslyoperating repetitivelypulsed system at 4 kW of average power. The EHT solidstate power system can drive peak currents of ±2 kA at 500 kHz in the inductive RMF coils. UM integrated the power system with the thruster and conducted performance measurements in their Large Vacuum Test Facility. With the EHT power system they demonstrated parametrical control of the thruster performance (thrust, specific impulse, and efficiency) by varying the plasma input energy, neutral flow rate, and applied magnetic field. The new EHT power system and UM thruster upgrades improved the coupling efficiencies from ~3% to more than 50% and active specific impulses approaching 1000 s, while allowing for a direct thrust measurement. This is the first direct performance measurement of a continuously operating RMF–FRC Thruster using a standard invertedpendulum thrust stand in spacelike conditions. 

PP11.00032: Surface charging during disinfection by dielectric barrier discharges Harry Fetsch, Sophia Gershman, Yevgeny Raitses Dielectric barrier discharges (DBDs) have shown effectiveness at disinfecting surfaces due to reactive gas species, UV, electromagnetic fields, and fluxes of charged particles to the surface. A relatively less explored factor is the transfer of charge to the treated surface. Resulting field strengths on the order of 10kV/cm, depending on pulse shape and biological species, can cause ruptured membranes, intracellular damage, and apoptosis. Weaker fields, when not directly damaging, can increase uptake of chemicals from the environment due to membrane electroporation. Plasma treatment provides cytotoxic species, making this a promising avenue for disinfection. In this work, we use a noncontacting voltmeter to measure the transfer of surface charge in treatment by two devices: a flexible printedcircuit device (flexDBD) and a commerciallyavailable floatingelectrode D’Arsonval device. Preliminary results for polyimide film suggest a field of up to 11kV/cm posttreatment. Information on surface charging is valuable in designing novel devices and operating regimes to optimize disinfection. 

PP11.00033: Spectroscopic Characterization of LowTemperature Plasma Produced by Intense Electron Beams A. Stephen S Richardson, Eric R Kaiser, Stuart L Jackson, David D Hinshelwood, Stephen B Swanekamp, Paul E Adamson, Michael J Johnson, Darryl J Watkins, Nancy D Isner The lowtemperature plasma produced by the electron beam emitted from a Febetron pulsedpower generator is being characterized spectroscopically. The generator located at the Naval Research Laboratory produces a peak voltage of 80 kV, peak current of 4 kA with a pulse width of 100 ns. This beam passes through a thin anode foil into a chamber filled with various background gases at pressures between 10 mTorr and 10 Torr. Emission spectra from the nearUV to visible (300650 nm) of the beamgenerated plasma are measured using a timeintegrated survey spectrometer, a 0.5 m focal length midresolution spectrometer, and a subangstrom resolution 1.3 m focal length spectrometer. The 1.3 m spectrometer takes multiple simultaneous spectral measurements per pulse, enabling measurements of spatial variations of the plasma spectra through a 7fiber input array. Imaging of the spectra is conducted by a 2D intensified gated CCD over small time increments to analyze the timeresolved evolution of the emissions. These measurements provide data on the population dynamics of N_{2} rotational states, which provides a useful validation dataset for modeling performed using recently computed statetostate cross sections for rotational states of neutral and ionized N_{2}. 
Not Participating 
PP11.00034: Secondary Electron Emission Measurements of Ionic Liquids Angela M Capece, Pierce E Wickenden, Nicholas M Lipari We present measurements of the secondary electron emission (SEE) yields of ionic liquids for primary electron beam energies of 50 to 1000 eV. Ionic liquids are salts that consist of an organic cation and polyatomic anion and are liquid below 100°C and thermally stable up to 450°C. As a result of their low vapor pressure, ionic liquids can be combined with lowpressure plasmas to produce metal nanoparticles through the reduction of dissolved metal salts by the reactive species produced in the plasma. Nanoparticles have been created with DC discharges using a pair of electrodes with one immersed in an ionic liquid containing dissolved metal precursors.^{1,2} In typical DC discharges, the plasma is sustained by SEE from the cathode and electronimpact ionization of the gas. For discharges operating with a liquid cathode, the processes leading to electron emission from the liquid are likely very different than from metal cathodes. In this work, we exploit the low vapor pressure of ionic liquids to make direct measurements of the SEE yield under vacuum by bombarding liquid samples of 1butyl3methylimidazolium tetrafluoroborate and 1butyl3methylimidazolium acetate with electrons and measuring the resulting currents. 

PP11.00035: Modelling the interaction of atomic boron with graphene surfaces Sierra E Jubin, Aaditya Rau, Aidan Harges, Yuri V Barsukov, Igor Kaganovich Adsorption can play an important role in materials processing applications, but care must be taken to properly address surface chemistry interactions in computational investigations. Here, we discuss the interactions of atomic boron with graphene sheets using density functional theory (DFT) calculations, investigating adsorption energies and geometries with a variety of functionals and basis sets. In addition, we investigate several potentials used in the molecular dynamics (MD) code LAMMPS, evaluating their ability to successfully replicate the interaction of atomic boron with graphene sheets observed using DFT calculations. Some pitfalls and their repercussions are discussed for larger scale simulations of atomic boron bombarding graphite. 

PP11.00036: Secondary Electron Emission from Complex Surfaces angelica ottaviano, Richard E Wirz Materials with complex surfaces have been shown to reduce secondary electron emission (SEE) yield because their microcavities can trap emitted electrons. These materials are useful for various plasma applications including fusion energy, electric propulsion, and manufacturing. Secondary electrons significantly affect the life of plasma facing surfaces and plasma performance due to their ability to modify the local sheath and cool the bulk plasma. In addition, SEE can contribute to plasma instabilities and anomalous crossfield currents. 

PP11.00037: Plasma Interactions and Sputtering Erosion for PlasmaInfused VolumetricallyComplex Surfaces Anirudh Thuppul, Mary F Konopliv, Richard E Wirz Plasma material interactions are a challenge for lifetime and performance of plasma devices for applications such as fusion energy and electric propulsion. Recent advances in the development of robust surfaces for electrodes and plasmafacing surfaces for high energydensity applications at the UCLA Plasma & Space Propulsion Laboratory have shown that volumetricallycomplex surfaces produce persistent sputtering yield reduction of 80% (Li et. al 2021). The objective of this work is to characterize the material sputtering behavior and bulk plasma response of plasma interactions with a range of volumetricallycomplex surface geometries. Both plasmainfused and plasmafacing regimes are investigated using flat aluminum and kapton surfaces, and 10 and 40 PPI aluminum foams with different backing materials exposed to a xenon plasma. The angular sputtering yield, optical emission line intensity, bulk plasma potential, density, electron temperature, sample current, and temperature are all constantly measured over the 15 hour exposures. The measurements, along with comparison between different PPI foams and backing materials, provide insight on the impact the effect the plasmafacing versus plasmainfused regime plays on persistent sputtering yield reduction, plasma properties, and contamination. 

PP11.00038: EDGE AND PEDSTAL,STELLARATORS


PP11.00039: Pedestal Physics Enabling High Fusion Performance and Low Recirculating Power for EXCITE and Pilot Plant Designs Philip B Snyder, John Canik, Jerry W Hughes, Matthias Knolker, Orso Meneghini, Tom H Osborne, Jin Myung Park, Wayne M Solomon, Theresa M Wilks, Howard R Wilson The pressure and temperature at the top of the pedestal play a key role in fusion performance, and complex physics within the narrow edge transport barrier regulates these. We employ and further develop the EPED model to predict the pedestal structure, and derive a set of metrics to evaluate pedestal contributions to performance. We review comparisons of EPED predictions to observations on several tokamaks, focusing on high pedestal regimes. Strong shaping and moderate aspect ratio facilitate operation with a high pressure pedestal limited by currentdriven kink/peeling modes ("peeling limited") even at relatively high density. In the peelinglimited regime, the pedestal is predicted not to be degraded by high separatrix density, facilitating compatibility with a dense radiative divertor plasma. Optimization of the pedestal facilitates not only high fusion power density but also very high (>80%) bootstrap current fraction, enabling compact devices with low recirculating power and continuous operation. A regime is identified with intermediate R/a~2.32.7, and strong shaping, which holds promise for nextgeneration fusion devices such as a compact fusion pilot plant (CFPP) and an Exhaust and Confinement Integration Tokamak Experiment (EXCITE). 

PP11.00040: Broadening the SOL Heat Load Width by Turbulence Spreading Patrick H Diamond, Xu Chu SOL widths in Hmode scale unfavorably with poloidal field. Since pedestal turbulence persists in Hmode, we calculate the turbulence intensity flux from the pedestal necessary to broaden the SOL beyond pessimistic HD predictions. A SOL width as a function of spreading flux is derived. In turn, the requisite intensity flux is related to pedestal turbulence levels and the transport barrier shear, so the conditions for SOL broadening are determined, explicitly. Turbulent pedestal states, such as “grassy ELMs”, wide pedestal QH mode, etc., are identified as desirable regimes, as they produce higher pedestal turbulence, which in turn broadens the SOL. Interesting implications for SOL stability in the case of strong layer broadening are discussed. 

PP11.00041: Scrapeoff layer transport from a gyrokinetic perspective Michael Churchill, Choongseok Chang, Seung Hoe Ku, Robert Hager Predictions of the divertor heatflux width on ITER is a critical issue which has major implications for design and operation. Simulations with the edge gyrokinetic turbulence code XGC1 predicted an ITER divertor heatflux width ~10x larger than the expected value based on scalings from current experimental devices [1] (Eich scaling). A plausible explanation for this was given that due to weaker ExB shearing layer (driven by device size), the turbulence correlation length is larger on ITER, and leads to turbulence spreading of the scrapeoff layer heat flux [1]. 

PP11.00042: Kinetic Study of Blob Dynamics in Realistic Geometry using the XGC1 Code Junyi Cheng, James R Myra, Scott E Parker, Seung Hoe Ku, Robert Hager, Choongseok Chang A largeamplitude fieldaligned coherent structure (a blob) is seeded as an initial condition in the closed fluxsurface region near the separatrix in the global gyrokinetic edge simulation code XGC1. XGC1 has realistic tokamak geometry including the scrapeoff layer (SOL) and divertor xpoint regions. In the simulations, a single blob is seeded with a uniform background density and temperature to simplify the physics and focus on the blob dynamics in the realistic magnetic geometry. Ions are gyrokinetic and electrons are driftkinetic. We are especially interested in how kinetic electron dynamics compare with fluid theory and simulation. We investigate the scaling of radial blob velocity versus the amplitude of the perturbation and compare it with estimates from fluid theory. Additionally, we find that, as the blob evolves, the structure is not strictly fieldaligned, and it appears to be crossing flux surfaces in a direction consistent with the ion magnetic drift. For large amplitude seeding, we see a strong spinning of the blob and we report the rotation frequency versus blob amplitude. The dipole field structure is seen for smaller perturbation strength but is washed out by the blob spin at a larger amplitude. The ExB drift still dominates the blob radial motion, since the averaged ExB drift agrees with the blob radial motion for both small and large amplitudes. We also investigate the radial blob motion and spinning versus collisionality. Finally, the theoretical description of the blob propagation for the current simulations will be presented. 

PP11.00043: Gyrokinetic Modeling of Electrostatic IonScale Turbulence in Divertor Tokamaks Mikhail Dorf, Milo Dorr, Debojyoti Ghosh, Ronald Cohen Continuum gyrokinetic simulations of electrostatic ion scale turbulence are presented for the case of a diverted tokamak geometry. The simulation model, implemented in the finitevolume code COGENT, solves the longwavelength limit of the fullF gyrokinetic equation for ion species coupled to the quasineutrality equation for electrostatic potential variations, where a fluid model is used for an electron response. The model describes the ionscale ion temperature gradient and resistive drift modes as well as neoclassical ion physics effects. Regimes of enhanced and suppressed turbulence are observed depending on the plasma profiles, and the roles of a selfconsistent background electric field and an Xpoint geometry are explored. In order to facilitate simulations of highlyanisotropic microturbulence, a numerical algorithm utilizing a locally fieldaligned multiblock coordinate system is employed in COGENT. In addition, the efficiency of numerical calculations is improved by making use of implicit time integration and the use of spatiallydependent velocity normalization that can facilitate simulations with large variations in a background plasma temperature. 

PP11.00044: Totalf Electromagnetic XGC in realistic diverted tokamak geometry Seung Hoe Ku, Robert Hager, Choongseok Chang, Amil Sharma, Michael Churchill The totalf gyro kinetic code XGC [1] has been making valuable discoveries across the magnetic separatrix in electrostatic modes, including LH bifurcation dynamics, blob dynamics, diverter heatload width physics, RMP physics, etc. We report the totalf upgrade of the XGC capability to electromagnetic physics for the first time, while retaining all the totalf capabilities that selfconsistently include neoclassical physics, turbulence physics, neutral particle recycling, heat and torque source, nonlinear FokkerPlanck collisions, logical sheath, etc. The totalf electromagnetic capability to be reported here is based on the reduced deltaf algorithm utilizing the mixedvariable and pullback transformation methods developed by Hatzky, Mishchenko,Kleiber et al., as described in Ref. [2]. Totalf electromagnetic algorithm and physically important totalf electromagnetic solutions will be presented. We have also developed an implicit XGC version [3], which will not be the subject of this presentation. 

PP11.00045: The gyrokinetic transport analysis version of the XGC code Albert V Mollen, Robert Hager, Choongseok Chang, Varis Carey, Nathaniel M Ferraro The upgrade of the totalf axisymmetric version of the gyrokinetic particleincell code XGC [R. Hager et al. Phys. Plasmas 26 104502 (2019)] to a gyrokinetic transport analysis code is reported. The code is equipped with anomalous 2D crossfield transport fluxes superposed to the Lagrangian particle motions. Neoclassical physics, neutral particle recycling and fully nonlinear FokkerPlanck collisions are retained together with the selfconsistent 2D axisymmetric electric field solution. In the XGC program, the driftkinetic XGC0 code has played the role of a transport analysis tool, but its capability has been limited to 1D fluxsurface averaged radial transport coefficients [D. J. Battaglia et al. Nucl. Fusion 53 113032 (2013)]. Unlike XGC0, the present code solves for poloidalangle dependent physics that is necessary for a more reliable description of the edge plasma in the pedestal, across the magnetic separatrix and in the scrapeoff layer. The 2D anomalous transport fluxes can be imported from a fullscale XGC simulation or from a surrogate model. Moreover, in the future we plan to use this version of XGC to telescope the plasma profile evolution through a multiscale coupling with a fullscale simulation, while the magnetic equilibrium is reconstructed by a code like M3DC1. 

PP11.00046: Gyrokinetic calculation of RMP penetration in a KSTAR Hmode plasma Robert Hager, SeungHoe Ku, Andreas Kleiner, CS Chang, Nathaniel M Ferraro, Jaehyun Lee The global totalf electromagnetic gyrokinetic code XGC is applied to calculate the penetration of external resonant magnetic perturbations (RMPs) in a KSTAR Hmode plasma. In contrast to conventional simulations of RMP penetration with magnetohydrodynamics (MHD), the XGC calculation retains the nonlinear kinetic interactions among the external RMP field, the screening currents in the plasma, toroidal rotation, crossfield transport, and neutral particle recycling. XGC has been successfully used to compute neoclassical and electrostatic turbulent transport with a static RMP field computed with M3DC1 [R. Hager et al., Phys. Plasmas 27, 062301 (2020)]. For the RMP penetration study, an electromagnetic totalf version of XGC [based on the deltaf version of M. Cole et al., Physics of Plasmas 28, 034501 (2021)] is used in realistic divertor geometry. The RMPs calculated with XGC are compared against MHD calculations with M3DC1 to assess differences between the kinetic and MHD plasma response. 

PP11.00047: Design of an innovative set of RMP coils for TCABR Felipe M Salvador, Gustavo P Canal, Dmitriy M Orlov, André Salgueiro Bouzan, Marlena N Kot An upgrade of the Tokamak à Chauffage Alfvén Brésilien (TCABR) is being carried out. TCABR is a small tokamak (R_{0} = 0.62 m, a ≤ 0.2 m, I_{p} ≤ 120 kA, B_{0} ≤ 1.1 T) operated at the Institute of Physics of the University of São Paulo, Brazil. Part of this upgrade consists in the installation of an innovative set of ELM control coils to allow for detailed studies of tokamak plasma response. This innovative coil set consists of three toroidal arrays on the low field side, composed of 18 invessel coils each, and three toroidal arrays on the high field side, composed of 18 invessel coils each. These coils are being designed using the M3DC^{1} code and they will allow for the creation of a wide range of magnetic perturbation geometries and spectra, with toroidal mode numbers up to n = 9. 

PP11.00048: Modeling of tungsten impurities with gyrokinetic bundles in the XGC code for the study of JETILW and WEST physics Julien Dominski, Aaron Scheinberg, Choongseok Chang, Jeronimo Garcia Olaya, Martin O'Mullane, Vassili Parail, Clarisse Bourdelle, Ahmed Diallo, Alberto Gallo Tokamak physics in a tungsten environment is a critical topic of research, as tungsten impurities are found to degrade the confinement time of tokamaks. This physics will be studied with the XGC code by simulating the whole device plasma with multiple gyrokinetic ions [1,2]. To take into account the many ionization states of tungsten, we have developed a new model that permits to model 50 ionization states of tungsten with only 7 gyrokinetic bundles and their atomic interactions [3]. Bundles are commonly used by fluid code, but not by gyrokinetic codes. The atomic interaction rates for the gyrokinetic bundles are computed from ADAS data. The motivation for including the atomic interactions comes from preliminary simulations where we observed that large (df/f~1) amplitude asymmetries, whose poloidal orientation depends on the charge number Z, form in the pedestal and edge regions. Furthermore, these different asymmetries drive particle fluxes of different directions and lead to the accumulation of tungsten in the pedestal. We will present our model and preliminary results for JETILW pedestal physics, as well as a new application of the XGC code to the study of WEST edge physics. 

PP11.00049: Imode impurity removal and energy confinement Silvia Espinosa Using highz wall materials attempts to switch the fusion challenge from heat load handling to removing impurities. We propose a means of measuring the radial impurity flux from currently available diagnostics, providing insight on optimal tokamak operation to prevent impurity accumulation [PoP 24, 055904 (2017)]. Our description is a modification of Per Helander's high Z impurity treatment [PoP 5, 3999 (1998)]. It uses poloidal impurity flow measurements rather than a main ion kinetic calculation of screening effects. 

PP11.00050: TransportDriven Toroidal Rotation with Generally Varying Diffusivity Timothy J StoltzfusDueck Future devices like ITER will have limited capacity to drive toroidal rotation, increasing the risk of instabilities like resistive wall modes. Fortunately, many experiments have found that tokamak plasmas rotate “intrinsically”, that is, without applied torque. The modulatedtransport model shows that such rotation may be caused by the interaction of ion driftorbit excursions with the strong spatial variation of the turbulent momentum diffusivity. The model captures intriguing qualitative behavior, such as the strong dependence of edge intrinsic toroidal rotation on the majorradial position of the Xpoint. However, quantitative modeling applications need more detailed experimental features. For example, the original model required the turbulent momentum diffusivity to decay exponentially in the radial direction, while experiments have more complicated variation. In this work, we generalize the modulatedtransport model to allow the turbulent momentum diffusivity to depend on space in a completely arbitrary way. The normalized diffusivity is assumed to be weak, a condition that is typically met for experimental applications, because the normalization is large. The new calculation may serve as a basis for future extensions, including shaped geometry and trapped particles. 

PP11.00051: Numerical considerations for simulating the Hmode physics using gyrokinetic codes W. W. W Lee


PP11.00052: Eliminating FiniteGrid Instabilities in Gyrokinetic ParticleinCell Simulations Benjamin J Sturdevant, Luis Chacon Finitegrid (aliasing) instabilities place severe limitations on momentumconserving particleincell (PIC) methods applied to models with charge separation effects by requiring the resolution of the Debye length. Gyrokinetic models, on the other hand, enforce quasineutrality thereby removing the Debye length analytically. Recent studies with momentumconserving PIC applied to gyrokinetic models, however, show that a manifestation of this instability exists in certain physical parameter regimes for arbitrary spatial resolution [1,2]. Here, we show that a simple reformulation of the discrete equations, using a colocated discretization of the continuity equation, eliminates this instability for all practical purposes. We perform numerical dispersion analyses for both the original and reformulated schemes, including the effects of finite drifts and finite beta. This reformulation may be useful for codes with complicated meshes, where highorder shape functions or energyconserving schemes are difficult to implement. Finally, we demonstrate that our reformulation eliminates the finitegrid instability in an implicit electromagnetic version of the XGC code. 

PP11.00053: Development of an Unstructured Mesh Based Gyrokinetic Particleincell Code for Exascale Fusion Plasma Simulations Chonglin Zhang, Gerrett Diamond, Cameron W Smith, Mark S Shephard Particleincell (PIC) methods are being widely used for fusion plasma simulations. Unstructured meshes are a natural choice for the domain discretization due to the complexity of the simulation domain. In this poster, we discuss the development of a new distributed unstructured mesh gyrokinetic PIC code named XGCm, short for xpoint included gyrokinetic code meshbased. The code adopts the physical algorithms from the wellestablished XGC code. It is developed on top of the Kokkos based PUMIPic library, which is a distributed unstructured mesh infrastructure for PIC simulations. The code is developed to run on current GPU hardware, and is aimed at exascale fusion plasma simulations. First, the numerical algorithms suitable for distributed unstructured mesh and achieving scalability on GPU devices will be presented. Extensive unit tests and integration tests were created for code verification and will be presented. Code validation will be presented using the cyclone base case (case 5 from Burckel, etc. Journal of Physics: Conference Series 260, 2010, 012006). The turbulence growth rate will be compared with existing results to validate the overall code. Additional code validations will be included. Finally, Summit scaling results will be presented. 

PP11.00054: A kinetic LandauFluid closure for cutoff Maxwellian distribution Kaixuan Fan, Xueqiao Xu, Ben Zhu, Pengfei Li A new kinetic LandauFluid closure, based on the arbitrary truncated Maxwellian distribution is derived. A special case is considered in the static limit (the frequency ). This analytical closure recovers the kinetic LandauFluid closure when the cutoff velocity in the truncated Maxwellian distribution function is infinity. The gradient of perturbed heat flux ∂q/∂z is related with both ∂T/∂z and ∂^{2}T/∂z^{2}. We find an additional heat flux caused by the average velocity of particles when the truncated Maxwellian distribution is asymmetric. A physics explanation of the closure will be discussed. It is a general LandauFluid closure for fluid moment models, that suits a particles distribution in an open magnetic field line region, such as the ScapeOffLayer (SOL) of tokamak plasmas. 

PP11.00055: Innovative NonResonant Divertors Applied to Compact Toroidal Hybrid (CTH) Kelly Garcia, Aaron C Bader, Gregory J Hartwell, John C Schmitt, Oliver Schmitz Nonresonant divertors separate the confined plasma from surrounding structures with the resulting boundary region comprised of cantori and/or stochastic regions, but without the presence of large islands. In contrast, island divertor configurations make use of low order rational surfaces with large islands mediating the confined plasma and the wall. These islands are highly sensitive to the value and shear of the rotational transform which can be affected by the evolution of the plasma equilibrium. Compact Toroidal Hybrid (CTH) can serve as a testbed for the nonresonant divertor solution for divertor optimization. The currents in the field coil and ohmic current drive systems of CTH are controlled to alter the rotational transform between 0.3 < ι < 0.75. Utilizing the FLARE fieldline following code, we calculate strike point locations for the exiting plasma for multiple ohmic current values. These calculations provide possible locations for divertor plates that will be built in the experiment to test nonresonant divertor resiliencies. These same techniques can be applied to other machines including ones that use the island divertor in the standard operation, like W7X. 

PP11.00056: Density Limit Disruption Studies in the Compact Toroidal Hybrid Experiment using Machine Learning Based Bolometer Inversions James Kring, Stephen Knowlton, David A Maurer, David A Ennis, Gregory J Hartwell Recent bolometer data from density limit disruption studies in the Compact Toroidal Hybrid (CTH) at Auburn University show spatial and temporal correlations with poloidal magnetic field fluctuations. Using lineofsight approximations, the source of the bolometer fluctuations appears to lie on the reconstructed flux surface. Using a DeConvolutional Neural Network to inverted the bolometer data, the inverted bolometer fluctuations rotate spatially with the measured magnetic field fluctuations which have the form of a MHD mode. This result points to the origination of the bolometer fluctuations to be on the Opoints of the 2/1 mode. The intensity of the bolometer and poloidal magnetic field fluctuations grow synchronously and rapidly just before the density limit disruption. 

PP11.00057: Calibration of a Hall Probe Array for Equilibrium Reconstructions in the Compact Toroidal Hybrid Noah P Bessard, David A Ennis, David A Maurer, Gregory J Hartwell An array of 16 gallium arsenide Hall sensors spanning 121.5 mm has been implemented as a magnetic field diagnostic in the CTH device. The main advantage of using these sensors is their 2.25 surface area which allows for local field measurements. Local magnetic measurements are necessary for V3FIT reconstructions of the poloidal magnetic field produced by plasma currents. Benchtop calibrations of the Hall probe array determined the sensors have an average sensitivity of 1.25 volts per tesla for fields perpendicular to the sensor surface and are generally insensitive to fields parallel to the surface. In situ calibration in the CTH device revealed a significant effect due to a slight difference in the angular orientation of the individual sensors in the array ranging from .8 to 1.9 degrees. The CTH vacuum vessel and coil frames are also known to produce a measurable flux due to eddy currents that arise from their mutual inductance with the CTH coils. When accounting for the sensitivity to orientation and the effects from eddy currents, the experimental magnetic field profiles are in good agreement with the modeled profiles from a BiotSavart code. 

PP11.00058: XRay production and mitigation on the CTH experiment Roger Dorris, David A Ennis, Gregory J Hartwell, David A Maurer The Compact Toroidal Hybrid experiment (CTH) at Auburn University has produced unexpectedly high XRay dosages in the surrounding facility. Two mechanisms for these emissions have been proposed: runaway electrons, driven by plasma current, and magnetic pumping. Here, we describe a measurement campaign where the XRay emissions are characterized by correlation experiments, using an XRay scintillator and a Geiger counter. A range of plasma conditions, involving various densities, rotational transforms, and accelerating potentials are produced, and the measured emission profiles are used to derive statistical distributions of XRay production. The observed energy, intensity, and spatial distributions of XRays are then correlated in time with the evolution of plasma conditions inside CTH, and are used to understand the XRay sources and develop mitigation strategies. 

PP11.00059: Spectral measurements of neutral density in plasmas with varying fractional ionization Eleanor N Williamson, David A Ennis, Gregory J Hartwell, Curtis A Johnson, Stuart D Loch, David A Maurer, Saikat Chakraborty Thakur, Edward E Thomas Understanding the transition region between fully ionized and neutrally dominated plasmas is important to the study of the magnetosphere of the earth, the corona/chromosphere transition regions of the sun, and detached divertors in fusion devices. Determining the fractional ionization of the plasma requires accurately measuring the neutral density. We use an absolutely calibrated spectrometer coupled with results from the collisional radiative solver ColRadPy to measure neutral density in two laboratory plasma experiments. Neutral density measurements are validated against pressure in the ALEXIS experiment an RF generated magnetized plasma column. Additionally, spectral measurements were collected from ECRH heated plasmas with varying fractional ionization in the Compact Toroidal Hybrid device. A triple probe and interferometer measure electron density ranging from 1 x 10^{17} m^{3} to 1 x 10^{18} m^{3} and temperature from 1 eV to 10 eV. Results will be presented from a study of fractional ionization within the plasma volume as well as the effects of metastables in both experiments. 

PP11.00060: Highresolution ultraviolet spectroscopy of tungsten and tantalum emission for improved erosion measurements Tomas Gonda, Curtis A Johnson, David A Maurer, Stuart D Loch, Gregory J Hartwell, Peter J Traverso, Nicholas R Allen, David Ennis Characterizing erosion of Plasma Facing Components (PFC) in magnetically confined experiments is necessary for understanding edge impurity dynamics and PFC lifespan in future fusion devices. A spectroscopic technique for temporally determining PFC erosion relates observed spectral line intensities to material influx via atomic physics coefficients (S/XB). However, failing to sufficiently resolve measured spectral lines due to impurity line blending, pressure broadening, and Zeeman and Stark splitting may lead to incorrect erosion estimations. A 1.3 meter focal length spectrometer is being optimized for highresolution (8 pm) ultraviolet (200400 nm) spectral measurements in the Compact Toroidal Hybrid (CTH) device. W and Ta samples are introduced into the plasma at varying depths with simultaneous Langmuir probe measurements of electron parameters. Comparisons of highlyresolved W spectra to recently computed atomic data for neutral W and W+ are made using the ColRadPy collisional radiative suite of codes to account for potential broadening mechanisms. First spectroscopic measurements of Ta emission in CTH are presented along with identifications of Ta spectral lines for potential benefit to erosion measurements. 

PP11.00061: Analysis of sawtooth simulations in the Compact Toroidal Hybrid (CTH) experiment Omar E Lopez, Eric Howell, James D Hanson, David A Maurer Sawtooth oscillations in tokamaks and currentcarrying stellarators display a gradual temperature rise in the core followed by a fast crash. In the Kadomtsev model [1], the crash is due to a linear mode that is triggered as the onaxis safety factor drops below unity. The Compact Toroidal Hybrid (CTH) device is a currentcarrying five field period torsatron in which a variable external rotational transform enables the study of the influence of 3D magnetic fields on MHD stability. Here, we analyze 3D nonlinear NIMROD sawtooth simulations for a CTHlike scenario [2]. A new fixed points finder implementation in NIMROD enhances the description of sawtooth cycles. Fixed points result from the intersection of a closed magnetic field line with a Poincare crosssection; they constitute the skeleton of the magnetic field structure. The implementation allows for the determination of Greene's residues [3], which can serve to estimate an island width and have also been used to measure the degree of stochasticity of stellarator vacuum magnetic fields [4]. In this work, the determination of order1 fixed points, Greene's residues, and local values for the rotational transform lead to a description of sawtooth oscillations in CTH consistent with the Kadomtsev model. Finally, to explore the distribution of energy among the Fourier modes and the quasilinear and nonlinear interactions during sawtooth cycles, we have expanded the Ho and Craddock [5] power transfer analysis to a 3D equilibrium. 

PP11.00062: Coordinate maps, ∇Ψ_{toroidal}, and B × ∇Ψ_{toroidal} vectors for 5/6, 5/5 and 5/4 island domains in W7X John C Schmitt, Matt Kriete, T. Andreeva, J. Geiger, M. Grahl, J. Schilling, H. Thomsen, E. Flom The edge island domain in Wendelstein 7X consists of divertor islands that coincide with the location of rational values of the rotational transform ι≈(5/6, 5/5, 5/4) and surround the main confinement volume (the ‘main plasma’). The "5/5" edge is 5 individual islands that are unconnected. In contrast, a single island connects onto itself after 6 or 4 toroidal transits in the the “5/6” and “5/4” edge, respectively. Many interesting phenomena are related to these islands and diagnostic analyses require a mapping from `laboratory' or real space coordinates to the island coordinate system. Two procedures are described here to calculate several scalar and vector quantities for closed island structures which can be utilized in fast interpolation schemes for inverse maps. For the "5/5" edge, a fixedboundary vacuum (zero beta) magnetohydrodynamic solutions of the 5/5 island is found with VMEC. The solution is compatible with already existing routines which determine ∇Ψ_{toroidal} (and other quantities) of VMEC solutions at arbitrary laboratory coordinates via stellarator symmetry. VMEC does not support solutions for the 5/4 and 5/6 island cases, so a fieldline following technique for calculating the ∇Ψ_{toroidal} and B × ∇Ψ_{toroidal} vectors along the field line is utilized. 

PP11.00063: STRAHL modeling of iron impurity transport with on and offaxis heating during the first divertor campaign on Wendelstein 7X Peter J Traverso, Novimir A Pablant, Andreas Langenberg, Thomas Wegner, Benedikt Geiger, Birger Buttenschön, Daihong Zhang, Håkan M Smith, James D Kring, John C Schmitt, Rainer Burhenn, Felix Reimold, David A Maurer In the first divertor campaign of Wendelstein 7X, iron impurity transport experiments were performed via laser blowoff injection during an on to offaxis ECRH scan at constant power. The observed iron spectral lines show an increase in the impurity transport time as a larger fraction of ECRH power was deposited offaxis and as the ECRH power was decreased. Although the purely onaxis power scan demonstrated a similar core Te flattening as the 4.9 MW offaxis scan, the resulting transport time enhancement was substantially larger for the onaxis power scan. To characterize these observed changes a least squares minimization was performed to infer the anomalous transport profiles that most accurately produces the measured iron line emission using the 1D transport code STRAHL. In all cases the observed iron line emission could only be wellmatched when the anomalous diffusion channel was included and at levels ~50 times larger than neoclassical predictions. A sensitivity study using synthetic data was performed capturing the systematic uncertainties in the inferred diffusion profiles. Although these profiles are consistent with an increase in the transport times across the onto offaxis dataset, the profiles are indistinguishable when these total uncertainties are considered. 

PP11.00064: Drift velocity and ion temperature measurements in the W7X scrapeoff layer using coherence imaging spectroscopy David M Kriete, John C Schmitt, Valeria Perseo, Dorothea Gradic, David A Ennis, David A Maurer, Ralf König To investigate island divertor physics, 2D ion velocity and temperature measurements are achieved in the scrapeoff layer (SOL) of W7X using coherence imaging spectroscopy (CIS). The CIS technique encodes information about lineintegrated ion velocity and temperature into a spatial interference pattern that is overlaid on an image of the plasma emissivity. Velocity measurements in plasmas with matched upstream parameters but opposite magnetic field directions show that in low edgeiota magnetic island configurations (ι = 5/6) drifts contribute substantially to SOL flows. The density scaling of the drift velocity varies throughout the SOL: in some regions it increases with density, while in others it decreases. Initial 2D ion temperature measurements derived from the contrast of CIS interference patterns are presented. Due to the high magnetic field of W7X (B_{0} = 2.5 T), Zeeman splitting substantially affects CIS contrast and must be accounted for to obtain accurate ion temperatures. A technique is described for estimating the Zeeman contrast, which varies across the CIS field of view. A design is also presented for a new CIS instrument simultaneously optimized for maximum sensitivity to ion temperature and minimum sensitivity to Zeeman splitting. 

PP11.00065: Overview of neoclassical flows and electric field studies in HSX Santhosh T Kumar, Joseph N Talmadge, Yasuhiro Yamamoto, Sadayoshi Murakami, Dimitrios N Michaelides, Konstantin M Likin, David T Anderson The reduced parallel viscous damping in the direction of symmetry for a quasihelically symmetric stellarator allows for large helical plasma rotation. Previous experiments in HSX have demonstrated good agreement with neoclassical modeling in terms of direction, and to some extent the magnitude, of plasma flows and neoclassical currents. Experiments and modeling are underway to further improve our understanding of flows and radial electric field in HSX. The effect of neutrals on the flow damping is being studied using multiple puffing locations, in conjunction with the neutral simulation code DEGAS. The neoclassical code PENTA has been modified to include collisions with background neutrals. New viewing port locations have been identified for Charge Exchange Spectroscopy to improve measurements of PfirschSchlüter flows. Feasibility of using an alkali neutral beam for measuring time evolution of flows is being studied. To validate GNET code simulations on the ECH driven parallel flows, measurements with Xmode ECH is planned. A biased electrode experiment is being carried out to understand the effect of parallel flow on the ion resonant electric field. Time dependent neoclassical models are being extended for low collisionality ions that are expected in the HSX upgrade. 

PP11.00066: A Correlation ECE Diagnostic for the HSXUpgrade Stellarator Luquant Singh, Konstantin M Likin, Jason Smoniewski Here we present a new correlation ECE diagnostic for the Helically Symmetric eXperiment Upgrade (HSXU) stellarator. HSX has demonstrated a significant reduction in neoclassical transport compared with conventional stellarators. Anomalous transport, likely due to Trapped Electron Mode turbulence, remains an important loss channel. For operation at 1.0T, a 16channel heterodyne ECE radiometer was developed for highresolution electron temperature measurements. However, this system is sensitive only to temperature fluctuations above ${\sim}2.5\%$, well above the level predicted by gyrokinetic simulations. Furthermore, we show that only relatively high$k$ fluctuations with poor radial localization are accessible with this conventional system. To measure turbulent temperature fluctuations at the level predicted in HSXU, a correlation ECE technique can be used. Based on electron temperature fluctuation and heat flux predictions from gyrokinetic simulations, we present a new correlation ECE diagnostic for HSXU. ECE optical depth is estimated assuming representative profiles for HSXU plasmas over a range of central electron temperature and plasma density and indicates blackbody emission. 

PP11.00067: Progress on the Upgrade of the Helically Symmetric Experiment Alexander L Thornton, Benedikt Geiger, David T Anderson, Konstantin M Likin The helically symmetric experiment (HSX) is an optimized stellarator using quasihelical symmetry (QHS) of the magnetic fields to confine its plasmas. It has been in operation since 2001 and has successfully demonstrated minimized neoclassical transport and relevant turbulence physics. HSX performance is limited by the frequency of its electron cyclotron heating (ECH) source, which does not allow plasma densities higher than 1x1019 m3. In order to triple this limit, HSX is upgrading its facilities to operate with a gyrotron recently acquired from the Max Planck Institute for Plasma Physics in Germany. All conceptual design work is now complete, and most engineering has been finalized. Fabrication, installation, and commissioning is now fully under way. Magnet power supplies have been upgraded and commissioned, and a new ECH power supply is currently being installed. Gyrotron and transmission line parts are being fabricated and installed. First plasma is expected by the end of 2021. 

PP11.00068: "Upgrade of the HSX Motor/Generator system for Operation at B=1.25 T" Syed Waris Mahmood, Paul A Willis, Alexander L Thornton, David T Anderson, Santhosh T Kumar HSX is being upgraded for high power ECRH operation at 70 GHz. The motor/generator system is modified to drive 50% more power to provide the needed magnetic field of B=1.25T. This results in increasing the speed of operation and the number of modular units in the system. The control system is being upgraded with a new control program and operating sequence of the motor/generator units to achieve the upgraded parameters. Upgrades also include temperature management and dual method speed measurement of the units for added safety and reliable operation. Vibration sensors have been installed with functionality of longterm spectral analysis to monitor bearing health for additional safety and to predict failure. Bearing defects manifest as impulses at frequencies related to bearing geometry. High harmonics are observed via filtering and envelope to increase SNR. Fault detection and alarm mechanisms are being added to increase system visibility and to minimize mean time to recovery. In addition, the safety system of the machine is being upgraded to a category 3 safety system. All modular units have been rebuilt, recommissioned and have been tested at upgraded parameters. Results of testing, design changes, and operation will be presented. 

PP11.00069: A vacuum bakeout system for the HSX stellarator upgrade Zander N Keith, Syed Mahmood, Alexander L Thornton, Paul A Willis, Kyle Roethle, David T Anderson Reducing neutral influx from the vessel wall is of critical importance for the higher power and higher density discharges of the HSX upgrade. A bakeout system has been designed for the HSX vacuum vessel, which is expected to pump out adsorbed gases, especially water vapor, from the stainlesssteel wall. A thermal simulation has been carried out in SolidWorks to determine heater layout, power requirements, and insulation choice such that the vessel is evenly heated up to 150C with wire wound heating strips. Each half period of the vessel is divided into three heating zones, with an additional zone on the closest boxport face. These zones will be separately controlled by a PLC system which will provide individual PID control to each zone to ensure uniform heating, which is crucial to the delicate weld joints of the vessel. The control loop will be closed by thermistors and will use pulse width modulation to maintain all zones at the same temperature. The PLC will also provide startup control sequencing to achieve uniform rate of temperature increase across all zones. 

PP11.00070: Towards realizing stellarator fusion: a new quasisymmetric stellarator experiment Chris C Hegna, Aaron C Bader, Benjamin Faber, David T Anderson, John Canik, Thomas G Kruger, Ian J McKinney, John C Schmitt, Paul W Terry, Benedikt Geiger, Oliver Schmitz Bolstered by recent advances in stellarator theory and computation [1] and the successes of Wendelstein 7X, the stellarator is poised to play a more prominent role in the magnetic fusion program. Recent theoretical advances are determining ways to use 3D shaping to improve energetic ion confinement and lower turbulent transport rates consistent with favorable neoclassical transport properties. Moreover, advances in optimization techniques have improved our ability to design new configurations and simplify coil design. Quasisymmetric stellarators have additional advantages including the presence of minimal neoclassical flow damping in the symmetry direction which potentially enables enhanced confinement regimes and more robust magnetic surface quality. Coupled with the intrinsic advantages of the stellarator including high density operation, availability of steady state and low disruptivity, quasisymmetric stellarators can be an ideal candidate for fusion power plants. In this work, we detail how these recent advantages can be realized in a new proposed stellarator experiment whose operation can be used to demonstrated improve stellarator confinement properties and test the viability of nonresonant divertors. 

PP11.00071: Coil Tolerances Motivated by Island Widths Thomas G Kruger, David T Anderson, Aaron C Bader, Chris C Hegna, Caoxiang Zhu, Stuart R Hudson, Matt Landreman, Alessandro Geraldini Islands are inherent to stellarator equilibria and can negatively affect confinement and stability. In this presentation, we present a methodology that allows for the elimination of internal magnetic islands in new stellarator configurations. We optimize the island's width using shape gradients of coils which also allows us to determine the coil tolerances. A relationship exists between the island width and the island poloidal flux, which is the magnetic flux between the island's O and X points. Variations of the island poloidal flux due to variations of the coil magnetic field is given and allows the island width to be minimized. We first calculate shape gradients of the coils for the magnetic field. This allows us to calculate shape gradients of the coils for the island width. Using coil magnetic field shape gradients, we can derive shape gradients of the coils for the island widths. Since we solve for the coil shape gradients, we can also solve for coil tolerances by setting a maximum allowable island width. Since we derive an analytic equation for the coil tolerances, we can maximize the coil tolerances which will greatly reduce machine cost. Shape gradients are evaluated for a new stellarator configuration called, WISTELLA. We demonstrate island width minimization and give coil tolerances. We also propose potential avenues for coil tolerance optimization. 

PP11.00072: Comparisons of Energetic Particle Confinement in Stellarator Configurations Aaron C Bader, David T Anderson, Michael Drevlak, Benjamin Faber, Chris C Hegna, Sophia A Henneberg, Matt Landreman, John C Schmitt, Yasuhiro Suzuki, Andrew S Ware Recent progress in stellarator optimization has identified configurations with very good energetic particle confinement. These configurations are typically found by optimizing either quasisymmetry, neoclassical metrics such as ε_{eff} and Γ_{c}, or both. However comparisons across configurations have typically been difficult to obtain due to differences in device size and field strength. In this poster, we present alpha particle confinement results from 9 stellarator configurations of different types, including quasihelically symmetric, quasiaxisymmetric, quasiisodynamic, and σoptimized configurations. Results for calculations with and without collisions are presented. Analysis of the metrics shows that good alpha particle confinement is highly corellated with the Γ_{c} metric, while the correlation with quasisymmetry is weaker. When collisions are included, configurations with losses near the trapped boundary perform relatively worse compared to configurations with deeply trapped losses. This suggests possibilities of optimizing for specific classes of collisionless particle losses. 

PP11.00073: Feasibility of a Fully Catalyzed DD Fusion Reactor Shawn Simko, Benedikt Geiger Previous studies have shown that a fully catalyzed deuteriumdeuterium fusion reactor greatly improves the power balance over a pure DD reactor, while removing the need for tritium breeding and reducing the average incident neutron energy found in a deuteriumtritium reactor. However, the greater temperatures required for sufficient fusion rates to obtain ignition require the consideration of relativistic effects and radiative loss mechanisms that are insignificant in a DT reactor. A model is found for the Lawson criterion of a generalized reactor accounting for relativistic effects, ee bremsstrahlung, synchrotron radiation, and first order corrections across a range of electron densities, temperatures, magnetic field strengths, and particle to energy confinement time ratios. The model is accurate in the temperature range ~10^{7}10^{9 }K and can be extended considering an expansion of bremsstrahlung to a greater number of multipoles. The reactor parameters for a stellarator using ISSO 4 v 3 scaling law are found, and optimum operating parameters are calculated. 

PP11.00074: On the measurement of bootstrap and PfirschSchlüter induced magnetic fields in stellarators using a beam based diagnostic Thomas P Crowley, Diane R Demers, Peter J Fimognari The magnetic topology of stellarators is altered by bootstrap and PfirschSchlüter currents in ways that can affect neoclassical transport, stability, and divertor performance. Simulation codes, constrained using pressure profiles and external magnetic measurements, are typically used to solve for the equilibrium and infer these currents. Our new technique uses a beam of alkali ions or atoms as test particles to probe the plasmainduced magnetic fields. The beam undergoes collisions with plasma particles, creating a spray of secondary ions that are detected outside of the plasma. The secondary ion momentum depends on the local magnetic vector potential at the beam particle's point of ionization and also along its path. We are investigating the relative importance of these two contributions and the feasibility of this technique for stellarators by simulating particles passing through the magnetic equilibrium of the Helically Symmetric eXperiment (HSX) and Wendelstein 7X. We have also developed a prototype detector that's been deployed on HSX to study particle and radiative noise signals that may impact diagnostic measurements. 

PP11.00075: A phenomenological model of nonambipolar particle transport to understand the SOL currents in Wendelstein 7X (W7X) Arun Pandey, Michael Endler, Felix Reimold, Matthias Otte, Lukas Rudischhauser, Valeria Perseo In the typical standard magnetic field configuration of W7X (edge iota = 5/5), it is observed that in attached plasma condition, the divertor receives a significant nonambipolar particle flux as parallel current in the long connection length region. In contrast, the currents become very small when the plasma is detached. Two arrays of 10 flush mounted Langmuir probes are used for collecting these currents in the lower and upper divertors (10 at each location). The spatiotemporal profiles of these parallel currents are not explained by only considering the thermoelectric current contribution. Therefore, it becomes imperative to take into account the contribution from the nonambipolar perpendicular drifts in the SOL, hence opening the possibility to understand the nonambipolar SOL drifts by analyzing the SOL currents. A phenomenological model has been developed to understand the contribution of perpendicular fluid drifts to the parallel nonambipolar fluxes to the target. Due to the long connection lengths and small pitch angles of the magnetic field lines in the boundary islands of W7X, even a gentler parallel pressure gradient leads to a steep poloidal pressure gradient, which in combination with the radial pressure gradient in the SOL, drive the perpendicular diamagnetic current. The total perpendicular diamagnetic current is calculated by the model. A nonvanishing perpendicular divergence of the diamagnetic current would force a parallel current towards the strike point on the target to make the total current divergence free. 

PP11.00076: Recent Advances on The He/Ne beam diagnostic for line ratio spectroscopy in the Island Divertor of Wendelstein 7X Erik R Flom, Oliver Schmitz, Tullio Barbui, Marcin Jakubowski, Frederik Henke, Maciej Krychowiak, Ralf Koenig, Stuart D Loch, Jorge M MunozBurgos, John C Schmitt A lineratio spectroscopy system based on thermal helium and neon collisional radiative models (CRM) has been implemented and successfully shown to enable measurement of n_{e} and T_{e} above two magnetically connected divertor targets in the standard divertor configuration of the Wendelstein 7X optimized stellarator. Spectral line emission from gas injection is channeled to multiple CzernyTurner spectrometers, allowing high spectral resolution measurements of diagnostic helium lines. Neon has also been implemented in select discharges to investigate the expansion of the measurement envelope of the diagnostic into the low T_{e} and high n_{e} regime found in detached discharges. 

PP11.00077: Diagnostic design for beam emission spectroscopy at Wendelstein 7X for ionscale turbulence measurements David R Smith, George McKee, Benedikt Geiger, Marcus G Burke, Jurgen Baldzuhn, Oliver Ford, Olaf Grulke, Peter Poloskei, Adrian von Stechow, Thomas Windisch We report on a feasibility study to perform fluctuation beam emission spectroscopy (BES) measurements at Wendelstein 7X (W7X) to observe ionscale turbulence. W7X is a neoclassicallyoptimized stellarator, and initial results indicate that turbulence transport plays a significant role in heat and particle confinement at r/a ≈ 0.7. BES measurements can provide 2D imaging of plasma turbulence on relevant spatial and time scales by observing the localized Dopplershifted Balmeralpha emission (n=3→2) from neutral heating beams. The constraints of fieldaligned sightlines at the observation volume, sufficient Doppler shift to isolate the beam emission manifold and high optical throughput are challenging to achieve with the nearly radial heating beams on W7X. We report on BES measurement configurations that satisfy the constraints and provide 2D coverage at ionscales in the edge and midradius regions. In addition, we report on diagnostic performance factors such as spatial resolution, kspace coverage, emission spectra, and signaltonoise estimates with highthroughput collection optics and highspeed, lownoise BES detectors. Finally, we layout plans to deploy a BES diagnostic system for ionscale turbulence measurements at W7X. 

PP11.00078: Gaussian process tomography of the effective ion charge Z_{eff} from multiple lineintegrated bremsstrahlung spectra Sehyun Kwak, Uwe Hergenhahn, Udo Höfel, Maciej Krychowiak, Andrea Pavone, Jakob Svensson, Oliver Ford, Ralf König, Sergey Bozhenkov, Golo Fuchert, Ekkehard Pasch, Kai Jakob Brunner, Jens Knauer, Petra Kornejew, Humberto Trimiño Mora, Thomas S Pedersen In magnetically confined fusion plasmas, the effective ion charge Z_{eff}, which is useful for determining impurity contaminations, power losses and transport, can be inferred from the plasma electronion bremsstrahlung, given the electron density n_{e} and temperature T_{e}. At the Wendelstein 7X stellarator experiment, visible and infrared spectrometers collect the plasma bremsstrahlung spectra along multiple lines of sight which provide information on the spatial distribution of Z_{eff} over the plasma. To infer spatially resolved Z_{eff} profiles, a Bayesian model has been developed within the Minerva framework. Z_{eff}, n_{e} and T_{e} profiles are modelled as Gaussian processes whose hyperparameters constrain their smoothness. These profiles are mapped to Cartesian coordinates by assuming that physical quantities are constant on the poloidal magnetic flux surfaces. Given Z_{eff}, n_{e} and T_{e}, the predictive (forward) model, which takes into account the calibration of spectral responses and reflected stray lights, predicts the multiple lineintegrated bremsstrahlung spectra observed by the visible and infrared spectrometers. Besides the spectrometers, the model additionally includes the interferometer and Thomson scattering system to infer n_{e} and T_{e}. The inferred Z_{eff}, n_{e} and T_{e} profiles are provided as the samples drawn from their posterior probability distribution. The smoothness (hyperparameters) of the profiles is determined by evidence optimisation based on the principle of Occam’s razor. 

PP11.00079: Gyrokinetic simulations of the helically trapped electron mode in the W7X stellarator Javier H Nicolau, Zhihong Lin, Gyungjin Choi, Pengfei Liu, Xishuo Wei, Guillaume R Brochard The first global gyrokinetic simulations of the helicallytrapped electron mode (HTEM) in the Wendelstein 7X stellarator are presented. Using the GTC code, simulations with a 3D equilibirum, kinetic electrons and a prescribed density gradient exhibit an unconventional trapped electron mode. The eigenmode extends along the field lines in the weak magnetic field region and shows a strong variation in the toroidal direction as it was the case in ion temperature gradient (ITG) simulations. However, HTEM is located in the inner side of the torus where the curvature becomes unfavorable in the socalled 'straight' section in W7X. The HTEM is excited by helicallytrapped electrons due to the W7X magnetic configuration. In contrast to tokamaks, HTEM propagates poloidally in the ion diamagnetic direction. Further nonlinear simulations show that the zonal flows are a subdominant HTEM saturation mechanism. An inverse cascade in toroidal harmonics is observed during saturation which is enhanced by the excitation of lown harmonics. HTEM can cause significant particle transport comparable to heat transport in ITG simulations with a similar normalized temperature gradient. 

PP11.00080: Successful FIDA diagnosis of fast ions from negativeion neutral beams near their injection energy Wataru H Hayashi, Daniel J Lin, William W Heidbrink, Christopher M Muscatello, Yutaka Fujiwara, Hiroyuki Yamaguchi Although chargeexchange cross sections peak at relatively low energies (<100 keV), judiciously chosen sightlines that reduce the relative velocity between the injected neutrals and the fast ions can measure highenergy ions. New sightlines that use a negativeion neutralbeam injection (NNBI) source as the active beam were installed at the Large Helical Device (LHD). Fastion Dalpha (FIDA) spectra of dominantly NNBinjected plasmas collected during the 22^{nd} LHD campaign resemble theoretical predictions. Data from the views are compared with standard views that use positiveion neutral beams as the active source. The comparison confirms that diagnostic exploitation of the relativevelocity dependence provides a means to measure highenergy ions produce by the NNBI. Data during the hydrogen and deuterium operations are also compared, as well as the dependence of the signals on electron density. 

PP11.00081: Raytrace correction and in situ calibration for collective Thomson scattering diagnostics in the LHD Masaki Nishiura, Shun Adachi, Kenji Tanaka, Shin Kubo, Naoki Kenmochi, Ryoma Yanai, Takashi Shimozuma, Teruo Saito A fusion reaction produces energetic charged particles in plasmas for selfburning. The confinement of charged particles determines the performance of thermonuclear reactors. Collective Thomson scattering (CTS) diagnostic is one of the candidates to study the transport of MeV charged particles. In the large helical device (LHD), we have developed the CTS diagnostic using a millimeter wave of 77 GHz, 154 GHz, and 300 GHz. A heterodyne receiver detects scattered radiation at an overlap location between a probing beam and a receiving beam. The receiver is usually calibrated absolutely by using a liquid nitrogen source or a blackbody source. The radiation intensities are quite low compared with an intensity level of electron cyclotron emission (ECE) under an actual measurement condition. Thus, we have combined ECE and an electron temperature measured by an incoherent Thomson scattering. In addition, a millimeter wave of 77 GHz starts refraction even though the electron density is far below the cutoff density. The tendency is more significant when the lines of sight for the probing and the receiving beams are oblique to the magnetic field due to a density gradient. Introducing a ray trace code corrects the error of measurement locations for the refraction, the Doppler effect, and the relativistic effect. The calibration is applied to an offtiming of the modulated probing beam for subtracting the background ECE when performing the CTS diagnostic. This novel method for bulk and fast ion diagnostics is powerful and valuable under reactorrelevant experiments. 

PP11.00082: Novel optimized stellarator configurations Matt Landreman, Bharat K Medasani, Caoxiang Zhu Several new stellarator configurations are presented, generated using the new stellarator optimization framework SIMSOPT. It is shown that both quasiaxisymmetry (QA) and quasihelical (QH) symmetry can be obtained to higher precision than demonstrated previously, both on a single flux surface and throughout a volume. A number of new QA configurations are presented that have much improved confinement of fusionproduced alpha particles compared to previous QA configurations. Finally, a method is demonstrated to optimize a stellarator's geometry to eliminate magnetic islands and achieve other desired physics properties at the same time. In this approach, two equilibrium calculations are run at each iteration of the optimization: one that enforces the existence of magnetic surfaces (VMEC), and one that does not (SPEC), with island residues penalized in the objective function to bring the two magnetic field representations into agreement during the optimization. 

PP11.00083: SIMSOPT: A python/C++ framework for stellarator optimization Bharat K Medasani, Matt Landreman, Florian Wechsung, Andrew Giuliani, Rogerio Jorge, Caoxiang Zhu We present a new software framework called SIMSOPT for stellarator optimization. SIMSOPT is a mixed language software framework written in python and C++ for flexibility and efficiency. It provides objectoriented programming tools for defining objective functions and parameter spaces for stellarator optimization. SIMSOPT partitions the objective function using a graphbased approach and allows for dynamic alteration of the problem size. External MHD codes such as VMEC and SPEC can be used in defining the objective function. SIMSOPT provides classes for geometric objects that are important for stellarators such as surfaces and curves in different representations. It also makes multiple magnetic field representations available. An efficient implementation of the BiotSavart law, including derivatives is available. It provides MPI based tools for parallelized finitedifference gradient calculations. Modern software engineering practices such as a large suite of unit tests and continuous integration are employed in the development of SIMSOPT. It is distributed through multiple channels including conda, docker, and python wheels. 

PP11.00084: Adjoint methods for quasisymmetry of vacuum fields on a surface Richard Nies, Elizabeth J Paul, Stuart R Hudson, Amitava Bhattacharjee Adjoint methods, recently introduced to the field of stellarator optimisation [13], lead to significant speedup by providing gradient information, avoiding slower gradientfree optimisation methods or expensive finitedifference evaluations of the gradient. We herein apply adjoint methods to vacuum magnetic fields, deriving the adjoint equations corresponding to two objective functions, targeting either a given rotational transform or quasisymmetry with a given helicity on a surface. To measure the deviation from quasisymmetry, a novel way of evaluating approximate flux coordinates on a single flux surface without the assumption of a neighbourhood of flux surfaces is proposed, as vacuum fields are not generally integrable. The shape gradients obtained from the adjoint formalism are evaluated numerically, checked against a finite difference evaluation and shown for several configurations. First applications to stellarator optimisation are also presented. 

PP11.00085: Fast particle optimization of a quasiaxisymmetric stellarator equilibrium using the Γ_{C} metric Alexandra LeViness, David A Gates, Kenneth C Hammond, Samuel A Lazerson, John C Schmitt An important goal of stellarator optimization is improvement of the confinement of fast particles, such as alpha particles created by DT fusion reactions. In this work, a fixedboundary, quasiaxisymmetric (QA) stellarator equilibrium was reoptimized using the method outlined in Bader et al—minimization of both deviation from quasisymmetry as well as the analytical quantity Γ_{C}, the latter of which represents the angle between magnetic flux surfaces and contours of J, the second adiabatic invariant [1]. This was done while also maintaining MHD stability and good neoclassical confinement. In addition, the alpha particle losses have been simulated and compared between the original and optimized equilibria in order to measure the effectiveness of this optimization strategy, which has not before been used for a QA stellarator. The starting equilibrium was the LI383 equilibrium, a 3period QA stellarator, scaled to the magnetic field strength of the proposed permanent magnet stellarator PM4STELL [2]. 

PP11.00086: A Comparison of VMEC and DESC 3D Equilibrium Codes Dario Panici, Daniel W Dudt, Rory Conlin, Egemen Kolemen 3D equilibrium codes are vital for stellarator design and operation, as they provide the base state for experimental plasma operation. Highaccuracy equilibria are also necessary for stability studies. Comparisons of two 3D equilibrium codes will be presented. VMEC, which uses a gradientdescent algorithm to reach a minimumenergy plasma state, and DESC, which minimizes the MHD force error in real space directly. Accuracy as measured by final plasma energy and satisfaction of MHD force balance, as well as other metrics, will be presented for each code. Differences between the results and the code methods will be discussed. 

PP11.00087: Improvements to the DESC code for finding and optimizing stellarator equilibria Rory Conlin, Daniel W Dudt, Dario Panici, Jonathan Schilling, Egemen Kolemen Abstract: DESC is a pseudospectral code that computes 3D MHD equilibria by directly solving the force balance equations JxB=grad(p). We present a number of recent improvements to the code for both finding and optimizing stellarator equilibria. Automatic differentiation allows fast and accurate computation of derivatives used for solving the force balance equations and determining sensitivity to inputs. These calculations are further accelerated by leveraging GPUs. Accurate derivatives are used in a perturbation method to explore the phase space of stellarators and efficiently compute high resolution equilibria. We also demonstrate computing free boundary equilibria with DESC. 

PP11.00088: Connections between quasisymmetric magnetic fields and anisotropic pressure equilibria in stellarators Eduardo Rodriguez, Amitava Bhattacharjee Quasisymmetric stellarators are an attractive alternative to axisymmetric magnetic confinement devices. They replicate the good neoclassical behaviour of the latter while enabling flexibility in their threedimensional design to avoid disruptionprone currents. However, until now, it has been accepted in the fusion community that global quasisymmetry in magnetostatic equilibria (which obeys j×B=▽p with p the isotropic pressure) is not achievable exactly in stellarators. As evidence, the GarrenBoozer overdetermination problem is often cited: construction of solutions by means of a nearaxis expansion break down beyond second order due to the appearence of overdetermined equations. We show that this limitation is not inherent to quasisymmetry, but rather a result of requiring it in ideal MHD force balance. Extending the equilibria to include asymmetric forces avoids the overdetermination problem. Anisotropic pressure plays a central role in this extension, as not only does it avoid the overdetermination problem, but it can be shown to also be derivable as an energy minimum from a variational principle with global constraints in the manner of Kruskal and Kulsrud or Grad. Numerical solutions obtained by nearaxis expansion are presented. The extended space of solutions opens the door to a more fundamental understanding of quasisymmetry and its practical construction for optimal design of future experiments. In particular, it suggests that there might exist solutions in MHD equilibria that are very close to exactly quasisymmetric. 

PP11.00089: Evaluating nonlinear turbulence saturation in quasihelically symmetric stellarator geometries Benjamin Faber, Aaron C Bader, Ian J McKinney, Joey M Duff, MJ Pueschel, Paul W Terry, Chris C Hegna A primary goal of stellarator optimization is to produce new configurations with improved 

PP11.00090: Toroidal electron and ion temperature gradient instability in stellarators Jason F Parisi, Felix I Parra, Michael Barnes, José Manuel García Regaña, Iván Calvo, Grzegorz Walkowski, Denis A StOnge, Michael R Hardman We study the properties of toroidal electron temperature gradient and toroidal ion temperature gradient modes in the plasma core of W7X and LHD. Because stellarators typically have a large aspect ratio, the ratio of the major radius to the temperature gradient length scale can be as large as the steep gradient region in tokamak pedestals and transport barriers. This results in linear modes that have a perpendicular wavenumber that is much larger than their wavenumber in the binormal direction, and are challenging to resolve numerically. Using the gyrokinetic code \texttt{stella}, we find these toroidal modes are the fastest growing modes at a wide range of perpendicular scales. 

PP11.00091: Optimization of stellarator geometry for low transport from ITG/TEM modes Deepesh B Verma, Michael T Kotschenreuther, Michael C Zarnstorff, Swadesh M Mahajan, David R Hatch Gyrokinetic analysis advances indicate that the transport barriers can arise because of the existence of a regime of strongly weakened ITG/TEM instabilities [1]. It should be possible to optimize the geometry of a stellarator to be in this regime. Both the Wendelstein 7X geometry and the National Compact Stellarator eXperiment (NCSX) geometry have been found to lie in this regime, and many transport barriers in tokamaks also lie in it [1]. Here, a very fast implementation of a Simplified Kinetic Model (SKiM) [1] is being developed for incorporation into the STELLOPT optimization code [2]. The SKiM often gives an acceptably accurate estimate of the ITG/TEM growth rate, but is many orders of magnitude faster than gyrokinetic simulations. Together with optimization of neoclassical transport, this can be used to develop stellarator geometries with exceptionally high energy confinement. 

PP11.00092: Initial Results of NIMSTELL, the Stellarator Variant of NIMROD Carl R Sovinec, Brian S Cornille NIMSTELL is a substantial refactoring of the NIMROD code to make nonlinear stellarator MHD computations efficient. The changes include 1) expanding geometric information in finite Fourier series over a generalized toroidal angle and 2) using vector potential from the H(curl) space. Previous stellarator applications of the standard NIMROD code [for example, 13] suffered slow convergence of the Fourier expansion with uniform meshing over the geometric toroidal angle and were limited to configurations where the coils and plasma could be separated by a toroidally symmetric surface. The new developments for geometry are overcoming these challenges, and magnetic field is inherently divergencefree with the use of vector potential in H(curl). This presentation focuses on initial nonlinear MHD applications of NIMSTELL, including heating and MHD topological evolution in straight and toroidal configurations. The efficacy of the spectralelement/Fourier representation with approximately fluxaligned 3D meshing is demonstrated. [1] M. G. Schlutt, et al. NF 52, 103023 (2012); [2] N. A. Roberds, et al., PoP 23, 092513 (2016); [3] T. A. Bechtel, Stellarator Beta Limits with Extended MHD Modeling Using NIMROD, PhD Dissertation, UWMadison, 2021. 

PP11.00093: Heat transport as a measure of the effective nonintegrable volume Elizabeth J Paul, Stuart R Hudson, Per Helander Given the large anisotropy of transport processes in magnetized plasmas, the magnetic field structure can strongly impact the heat diffusion: magnetic surfaces and cantori form barriers to transport while chaotic layers and island structures can degrade confinement. When a small but finite amount of perpendicular diffusion is included, the structure of the magnetic field becomes less important, allowing finite pressure gradients to be supported across chaotic regions and island chains. We introduce a metric for the effective volume of nonintegrability based on the solution to the anisotropic heat diffusion equation. To validate this metric, we consider model fields with a single island chain and a strongly chaotic layer for which analytic predictions of the relative parallel and perpendicular transport can be made. We also analyze critically chaotic fields produced from different sets of perturbations, highlighting the impact of the mode number spectrum on the heat transport. We propose that this metric be used to assess the impact of nonintegrability on the heat transport in stellarator equilibria. 

PP11.00094: Application of the Shape Gradient and Hessian Matrix methods to compute the Sensitivities of Magnetic Islands to parameter perturbations in Permanent Magnet Stellarators Amelia Chambliss, Caoxiang Zhu, David A Gates, Tony Qian, Michael C Zarnstorff We applied methods for quantifying sensitivities of error fields to shape deviations in stellarator coils to novel permanent magnet stellarators. Using permanent magnets to produce stellarator fields offers a promising alternative to coil complexity. In the modular coil approach, small deviations in coil shape can have detrimental effects on particle and energy confinement. Permanent magnet stellarators are, however, also sensitive to source perturbations, and as a result, a detailed analysis of the possible field perturbations that can be caused by permanent magnets is necessary for device construction. We applied the shape gradient [Landreman & Paul, Nuclear Fusion 58(7), 076023 (2018)] and Hessian matrix methods [Zhu et al. Plasma Physics and Controlled Fusion, 60(5), 054016 (2018)] to permanent magnet stellarators to analyze the impact of perturbations to permanent magnet parameters including position, orientation, and magnetic moment magnitude on the production of error fields. We used the JAX automatic differentiation toolkit in Python to compute analytic derivatives of error field quantities. These methods were applied to study the sensitivity of the resonant perturbation metric, an indicator of the width of magnetic islands in the stellarator field. The shape gradient and Hessian matrix methods were applied to analyze the sensitivity of this error field quantity to perturbations of permanent magnet parameters in the MUSE permanent magnet stellarator and the PM4Stell permanent magnet project under development at PPPL. Implications for construction methods and precision requirements will be discussed. 

PP11.00095: Design of an arrangement of cubic magnets with discrete polarizations for a quasiaxisymmetric stellarator experiment Kenneth C Hammond, Caoxiang Zhu, Douglas Bishop, Amelia Chambliss, Keith Corrigan, David A Gates, Alexandra LeViness, Robert Lown, Robert Mercurio, Adam Rutkowski, John C Schmitt, Eric Stamper, Dennis Steward The usage of permanent magnet arrays to shape the confining magnetic fields of stellarators has been a topic of recent interest due to the potential of permanent magnets to reduce or eliminate the need for nonplanar coils. As a proofofconcept for this idea, we have developed a design for an arrangement of cubic permanent magnets that works in tandem with a set of planar toroidalfield coils to confine a quasiaxisymmetric plasma with a field on axis of 0.5 T. All of the magnets in the design are constrained to have identical geometry and one of three polarization types in order to reduce fabrication costs while still producing sufficient field accuracy. Their spatial layout was developed in tandem with a support structure and assembly scheme. We present some of the key considerations leading to the design, including the procedure for choosing the optimal polarization of each magnet from a discrete set, the choice of polarization types to be used, and the partitioning of the magnets into wedgelike modules that enable a radial insertion concept. 

PP11.00096: Error Correction Techniques and Analysis for the Princeton Permanent Magnet Stellarator Adam Rutkowski, Kenneth C Hammond, Caoxiang Zhu, Amelia Chambliss, David A Gates Stellarators offer a promising path towards fusion reactors, but their design and construction are complicated by stringent tolerance requirements on highly complex 3D coils. A potential way to simplify the engineering requirements for stellarators is to use simple planar toroidal field coils along with permanent magnet arrays to generate shaping fields. In order to ensure sufficient field accuracy while minimizing engineering complexity and system cost, new techniques are required to correct the field produced by the permanent magnet arrays to within requirements set by plasma physics. Attention is given to minimizing those errors that are important to the formation of islands, rather than enforcing strict global engineering tolerances. This work describes a novel correction method developed for this purpose. This method will be applied to the design and construction of a 1/2 period of the National Compact Stellarator Experiment (NCSX) equilibrium using permanent magnet arrays. Analysis techniques and initial results using the method for error correction on the proposed permanent magnet stellarator are presented. 

PP11.00097: Stellarator Fields without Stellarator Coils: MUSE a table top PM stellarator Tony Qian, Douglas Bishop, Amelia Chambliss, Arturo Dominguez, Christopher Pagano, Dominic Seidita, Michael C Zarnstorff, Caoxiang Zhu MUSE is a tabletop permanent magnet (PM) stellarator that uses simple circular coils, together with a PM array, to produce a stellarator vacuum equilibrium. The plasma configuration (R = 30 cm, B = 1.5 kG) is optimized to be quasiaxisymmetric. We found that our PM array produces free boundary VMEC equilibria with neoclassical transport metrics a factor of 10100 better than that of any constructed device. This might be due to the larger degrees of freedom available to PM compared to modular coils. From an engineering perspective, our PM array is composed of simple, identical magnets with equal magnetization all polarized axially. This simplifies assembly and procurement, but finding such a uniform distribution poses significant challenges for continuous optimization. We will discuss discrete optimization methods used to address these challenges. We also examine sensitivity of the PM array to assembly displacements and material irregularity. The extension of these methods to larger devices is also discussed. 

PP11.00098: Modifying the geometry of stellarator coils in order to minimize the intercoil electromagnetic forces Caira Anderson, Stuart R Hudson The threedimensional shaping of external currentcarrying coils needed to confine plasmas in stellarators can result in complicated coil geometry, making it difficult to design economically attractive fusion power plants based on the stellarator concept. We aim to lessen the intercoil electromagnetic forces by minimizing the magnetic field present outside of the plasma volume through modifying the geometry of the stellarator coils. The objective function is the sum of (i) the traditional quadratic flux passing through a given boundary, so that a desired plasma boundary is achieved, (ii) a coillength term, so that the coils do not become too long, and (iii) a magnetic energy term. Using the Julia coding language, we calculate these quantities as either surface or line integrals, and we use auto differentiation (AD) to compute the gradient of the objective function in terms of the stellarator coil geometry. The gradient information allows, for example, the steepest descent algorithm to be used to find the minimum. By minimizing the magnetic field outside of the plasma volume, we reduce the amount of energy needed to confine the plasma, and we potentially decrease intercoil repulsion. Thus the cost of constructing and operating fusion power plants is lowered. 

PP11.00099: Stellarator Confinement as Parametric Resonance Jeffrey M Heninger We present a novel intuitive description for how stellarators confine particles in a plasma. Typically, the magnetic field is plotted using a Poincaré section in a transverse plane, where we see nested flux surfaces. If we instead plot only the two components of the magnetic field within the transverse plane ( B_{r}(r,z) and B_{z}(r,z) ), the magnetic axis must be a saddle or degenerate because no other equilibrium satisfies ∇×B = 0 and ∇·B = 0. The nested flux surfaces are formed as a consequence of parametric resonance. As a magnetic field line transverses the stellarator toroidally, it experiences a rotating saddle, which stabilizes the otherwise unstable system. We analyze past and current stellarators to show that their magnetic fields exhibit parametric resonance. We hope that this new understanding allows us to use ideas from parametric resonance to help design better stellarators. 

PP11.00100: 3D calculation of magnetic turnstiles in nonresonant stellarator divertor Alkesh Punjabi, Allen H Boozer A very important result on the magnetic turnstiles in the nonresonant stellarator divertor [A. Punjabi and A. H. Boozer, Phys. Plasmas 27, 012503 (2020)] is obtained. The exiting and entering magnetic turnstiles of a family of magnetic turnstiles in the nonresonant stellarator divertor do not leave and enter the last confining surface at the same location but on the opposite sides of the last confining surface. This result is contrary to the conventional understanding of the turnstiles in Hamiltonian mechanics [R. S. Mackay, J. D. Meiss, and I. C. Percival, Physica D 13, 55 (1984); J. D. Meiss, Chaos 25, 097602 (2015)]. A method is developed to calculate the 3D structure of the magnetic turnstiles in the annulus between the last confining surface and the wall in topological torii. This method is used to calculate the 3D structure of the magnetic turnstiles in the nonresonant stellarator divertor. It is found that there are two families of magnetic turnstiles with probability exponents 9/5 and 9/4. The questions of the nature of continuous toroidal stripes and the secondary family having a negative probability exponent from our simulation of nonresonant stellarator divertor [A. Punjabi and A. H. Boozer, Phys. Plasmas 27, 012503 (2020)] are resolved. The toroidal stripes are a distinct family of turnstiles; and the secondary family with negative exponent are not magnetic turnstiles; they are the positions of the largest radial excursion of the last confining surface. The method can be useful in the designing and optimization of divertors in toroidal fusion plasmas. This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Fusion Energy Sciences under Awards DESC0020107 to Hampton University and DEFG0203ER54696 to Columbia University. This research used resources of the NERSC, supported by the Office of Science, US DOE, under Contract No. DEAC0205CH11231. 

PP11.00101: Validation of Impurity Ion Flows Measured by Coherence Imaging Spectroscopy in the Compact Toroidal Hybrid Nicholas R Allen, David Ennis, David M Kriete, Gregory J Hartwell, Curtis A Johnson, David A Maurer, Cameron M Samuell, Steven L Allen Twodimensional profiles of lineintegrated impurity emissivity and velocity in the Compact Toroidal Hybrid (CTH) experiment are obtained with Coherence Imaging Spectroscopy (CIS), a polarization interferometry technique. CIS ion velocities inferred in CTH are on the order of 5 km/s and have been validated by comparison with two dispersion spectrometers yielding good agreement to well within the CIS and spectrometer measurement error bars of 2 and 3 km/s, respectively, for multiple sightlines within the CIS field of view. The CISobserved flow response to a complete reversal of the magnetic field and current directions is limited, indicating ion flows in CTH are determined by other factors. CIS measurements utilizing currentdrawing probes in the plasma edge indicate a 5 km/s reversal of the edge ion flow. Ion flow measurements with varying probe currents and magnetic field configurations along with contributions from other parameters, such as external magnetic fields and thermal drifts, are presented. 

PP11.00102: Compact Toroidal Hybrid Research Progress and Plans David A Maurer The Compact Toroidal Hybrid (CTH) is a torsatron/tokamak hybrid. The main goals of the CTH experiment are to study disruptive behavior as a function of the applied 3D magnetic shaping, and to test and advance computational tools able to describe 3D MHD physics such as the V3FIT reconstruction code and NIMROD modeling of CTH. Recent density limit disruption and hard xray generation studies will be overviewed and their relevance to tokamaks and quasiaxisymmetric stellarators will be discussed. Ongoing diagnostic development for the experiment includes development of a new bolometer tomography technique, new Hall probe array measurements, new spectroscopic studies of W and Ta, and coherence imaging of plasma flows. CTH also serves as a test bed for diagnostic development for our collaborations on the larger facilities like DIIID and W7X. 

PP11.00103: BEAMS: COMPUTATIONAL, ANALYTICAL, MEASUREMENT, AND DIAGNOSTIC TECHNIQUES FOR LASERS AND BEAMS


PP11.00104: Achieving banded spectral selectivity through coded apertures without energy resolving detectors Matthew P Selwood, Chris Spindloe, Chris D Murphy Laserplasma xray sources have garnered interest from various communities due to their ability to generate high photonenergies from a small source size. The passive imaging of highenergy xrays and neutrons is also a useful diagnostic in laserdriven fusion as well as laboratory astrophysics experiments which aim to study small samples of transient electronpositron plasmas. 

PP11.00105: New developments in the OSIRIS simulation framework Ricardo A Fonseca, Pablo J Bilbao, Sarah E Chase, Fabio Cruz, Fabrizio Del Gaudio, Stephen Dilorio, Frederico Fiuza, Thomas Grismayer, Anton Helm, Roman Lee, Fei Li, Martin L Lindsey, Bertrand Martinez, Joshua J May, Kyle G Miller, Zan Nie, Miguel Pardal, Jacob R Pierce, Kevin Schoeffler, Adam R Tableman, Rui P Torres, Frank S Tsung, Marija Vranic, Han Wen, Benjamin J Winjum, Xinlu Xu, Viktor K Decyk, Warren B Mori, Luis O Silva The OSIRIS Electromagnetic particleincell (EMPIC) code is widely used in the numerical modeling of many kinetic plasma laboratory and astrophysical scenarios. In this work, we report on the new developments recently introduced into the framework. We address the implementation of new particle pushers and field solvers, that improve the accuracy of the PIC algorithm, especially for high field/high momenta situations and studying the evolution of particle spin, and also to deal with curvilinear coordinate systems. We report on our progress on our linear (particleparticle) Compton scattering and nuclear fusion modules, as well as improved boundary conditions for overdense plasmas. We present the new code features in terms of diagnostics, such as OpenPMD support, pressure tensor, and photon diagnostics for QED scenarios. Furthermore, we describe new features implemented in the Quasi3D geometry, in particular the inclusion of QED effects, external EM fields, and exotic laser beams. We also focus on the developments done in the General relativity module for modeling neutron star and black hole magnetospheres including strong gravitational fields. Finally, we present new developments aimed at including deep learningbased methods to model collisional processes. 

PP11.00106: Full Kinetic and Drift Kinetic Descriptions of Electrons Within MITLs Near a Load Mark H Hess, Evstati G Evstatiev In this study, we examine the dynamics of electrons in cylindrically symmetric magnetically insulated transmission lines (MITLs) near a load. Our analytical model of MITLs considers the electron motion in the presence of vacuum electric and magnetic fields for current drives that are similar to power flow experiments at the Sandia National Laboratories Z Pulsed Power Facility. Our study considers two types of MITLs, namely the radial MITL and a spherically curved MITL. We examine the motion of the electrons using both an exact Lagrangian/Hamiltonian framework as well as an approximate drift kinetic model that incorporates both ExB and grad B drift motion. In general, the drift kinetic model allows for fast calculations of electron dynamics, and yields excellent comparisons to the full kinetic model. Drift kinetic results can show some disagreement with the full kinetic results when the electric field gradient length becomes comparable to the Larmor radius of the electron. Additionally, we compare both of these models to fully electromagnetic and fully kinetic calculations using the code EMPIRE developed at Sandia National Laboratories. 
Not Participating 
PP11.00107: Modelling Chromatic Emittance Growth in Staged Plasma Wakefield Acceleration to 1 TeV using Nonlinear Transfer Matrices Alexander G Thomas, Daniel Seipt A framework for integrating transfer matrices with particleincell simulations is developed for TeV staging of plasma wakefield accelerators. Using nonlinear transfer matrices in terms up to ninth order in normalized energy spread $\sqrt{\langle\delta\gamma^2\rangle}$ and deriving a compact expression for the chromatic emittance growth in terms of the nonlinear matrix, plasma wakefield accelerating stages simulated using the threedimensional particleincell framework OSIRIS 4.0 were combined to model acceleration of an electron beam from 10 GeV to 1 TeV in 85 plasma stages of meter scalelength with long density ramps and connected by simple focusing lenses. In this calculation, we find that for relative energy spreads below $10^{3}$ and normalized emittance below mmmrad, the chromatic emittance growth can be minimal. The technique developed here may be useful for plasma collider design, and could be expanded to encompass nonlinear wake structures and include other degrees of freedom such as lepton spin. 

PP11.00108: Hybrid quantumclassical simulations of relativistic charged fluids Julien Zylberman, Nuno F Loureiro, Fabrice F Debbasch Extreme plasmas, which are both quantum and relativistic, are relevant in several contexts ranging from Astrophysics to Inertial Fusion. I will present the first steps in a research program aiming at simulating these plasmas on quantum computers. The approach is based on the Dirac equation and relies on two key results. First, a generalization of the socalled Madelung transform shows that the Dirac equation actually describes a quantum relativistic fluid of spin 1/2 particles. Second, the Dirac equation can be discretized into quantum walks, which are a standard tool of quantum computing and also constitute a universal quantum computational primitive. Our first results are simulations of shocks in extreme fluids immersed in a uniform electric field. The simulations have been first performed on a classical computer, but we have also developed a new quantumclassical hybrid algorithm tailormade for current Noisy IntermediateScale Quantum (NISQ) computers and run the simulations on IBM’s quantum processors. The next steps, including the introduction of selfconsistent electromagnetic fields, will be also discussed. 

PP11.00109: LASERPLASMA WAKEFEILD, BEAMPLASMA WAKEFILED, AND DIRECT LASER ACCELERATORS


PP11.00110: Effect of density gradient on direct laser acceleration Robert Babjak, Alex V Arefiev, Louise Willingale, Marija Vranic Electron acceleration has drawn significant interest in the plasma physics community during the last decades, motivated by the development of intense laser facilities. Among the many proposed acceleration schemes, direct laser acceleration (DLA) distinguishes itself for providing high charge electron beams ~100s of nC. Recently, a Gaussian plasma density profile was used to accelerate a 140 nC electron beam with energies up to 500 MeV using a moderatelyrelativistic laser pulse [1]. A varying density profile might have had a significant influence on the DLA process in the experiment. If this proves correct, this will provide an opportunity for optimizing future experiments for obtaining highenergy electron beams with moderate laser intensities. 

PP11.00111: Particle integrator for particleincell simulations of ultrahigh intensity laserplasma interactions Kavin Tangtartharakul, Guangye Chen, Alexey Arefiev Particleincell codes are the most widely used simulation tools for kinetic studies of ultraintense laserplasma interactions. As a benchmark problem, we examine the motion of a single electron in a plane electromagnetic wave. We show surprising deterioration of the numerical accuracy of the PIC algorithm with increasing normalized wave amplitude for typical timestep and grid sizes. Two significant sources of errors are identified: strong acceleration near stopping points and the temporal field interpolation. We propose adaptive electron subcycling coupled with a third order temporal interpolation of the magnetic field and electric field as an efficient remedy that dramatically improves the accuracy of the particle integrator. 

PP11.00112: Strong interplay between superluminosity and radiation friction during direct laser acceleration of electrons within a magnetic filament ILin Yeh, Kavin Tangtartharakul, Hans Rinderknecht, Louise Willingale, Alexey Arefiev Even though direct laser acceleration at ultrahigh intensities has been extensively studied, the impact of superluminosity on electron dynamics remains relatively unknown. The superluminosity (phase velocity larger than the speed of light) is unavoidable during laser propagation through a plasma and therefore must be taken into account. We have examined the direct laser acceleration of electrons within a static magnetic filament driven by a highintensity laser within a plasma. We studied the regime where the electrons also experience the force of radiation friction caused by the emission of electromagnetic radiation. We found that the interplay of superluminosity and radiation friction manifests as an attractor effect: the electrons with various initial energies reach roughly the same maximum energy and emit the same power in the form of gamma rays. The discovered effect is directly relevant to laserplasma interactions at highintensity multiPW laser facilities. 

PP11.00113: Design of Experiments to Study Relativistically Transparent Magnetic Filaments Using OMEGA EP Matthew A VanDusenGross, Hans Rinderknecht, Kathleen Weichman, David R Harding, Alexey Arefiev, Jarrod Williams, Alex Haid In relativistically transparent interactions, intense lasers can drive relativistic currents and azimuthal magnetic filaments in classically overdense plasmas. Simulations and theoretical models predict these relativistically transparent magnetic filaments can produce beams of highenergy photons. We designed experiments to study the efficiency of electron acceleration and xray production in nearcriticaldensity foamfilled microchannel targets when driven by the OMEGA EP laser. Using analytical scaling laws supported by 3D particleincell simulations, we optimized the laser and target design parameters for maximum photon energy and efficiency. By varying the density of the target, the characteristic energy and number of radiated photons can be controlled. We also finetuned diagnostic systems to experimentally test the predicted characteristics of these electron and photon beams. We will present initial designs for gasfilled microchannel targets developed to achieve nearcriticaldensity targets. 

PP11.00114: Mev to Gev Direct Forward and Backward Acceleration of Electrons from Undersense Plasma using a Radially Polarized UltraIntense Laser Focus Nour El Houda Hissi, Enam Chowdhury In this work we perform a computational investigation of the direct acceleration of electrons produced during ionization of undersense neon gas using tightly focused and radially polarized Petawattclass short pulse lasers with wavelength range from 0.8 to 2µm by numerically solving the relativistically invariant Lorentz equations with nonparaxial fields, incorporating semiclassical tunneling ionization and MonteCarlo type sampling of the focal volume. The accelerated electrons energy gain increases at longer laser wavelengths and GeV energies are reached for electrons ionized from the neon inner shell. Backward acceleration of electrons is observed for a range of initial positions and phases of ionization of neon charge states. This apparent counterintuitive phenomenon is directly linked to the radial polarization state of the incident laser beam that results in a strong longitudinal electric field (E_{z}) when tightly focused. Electrons ionized near the focal center at the phase when E_{z} is pointed toward the forward propagation direction experiences an initial push in the backward direction. A parametric study of the phenomenon by varying laser parameters will be presented. 

PP11.00115: A compact laserplasmabased scheme for the production of positron beams Davide Terzani, Ligia D Amorim, Carlo Benedetti, Stepan S Bulanov, Carl B Schroeder, Eric Esarey The production of highquality positron beams is an essential component of a future, plasmabased highenergy electronpositron collider. However, facilities where plasmabased acceleration of positron beams can be experimentally studied are lacking. We present a compact, laserbased scheme for the production of positron beams. Positrons are produced via pair decay of the Bremsstrahlung radiation generated when a multiGeV, laserplasma accelerated electron beam interacts with a highZ solid target. A realistic phasespace distribution for the positrons is obtained by modeling the electron beam interaction with the solid target using the Monte Carlo code Geant4. The positrons are then injected and accelerated into a second laserplasma accelerator stage, operating in the linear regime, through a transport line formed by a dipole magnet and an external focusing element. 

PP11.00116: Nonlinear laserdriven wakefields with arbitrary phase velocity Bernardo F Malaca, Jacob R Pierce, Miguel Pardal, Ricardo A Fonseca, Warren B Mori, John P Palastro, Dustin H Froula, Jorge Vieira, Dillon W Ramsey The flying focus technique can generate laser pulses where the group velocity can take any value – even superluminal – independently from its phase velocity. These pulses are generating significant interest as they can be used to enhance the acceleration in laserplasma accelerators (e.g. by enabling dephasingless acceleration in laser wakefields [J. Palastro et al, PRL 2021]) or to enhance the acceleration quality by avoiding dark currents (e.g. by using a superluminal driver [J. Palastro et al, PoP 2021]). Because these pulses promise to open new regimes in other areas, new techniques were developed and experimentally tested to generate laser pulses with arbitrary group velocities [e.g. H. Kondacki et al, Nature Comms 2019]. 

PP11.00117: Effects of Chromatic Aberration in a Dephasingless Laser Wakefield Accelerator Manfred Virgil Ambat, Robert Boni, Jessica Shaw, Philip Franke, Kyle R McMillen, Matthew A VanDusenGross, Hans Rinderknecht, Dillon W Ramsey, Tanner T Simpson, John P Palastro, SeungWhan Bahk, Jake Bromage, Dustin H Froula In laser wakefield accelerators, the ponderomotive force of an intense laser pulse propagating through a plasma excites a largeamplitude plasma wakefield that can trap and accelerate electrons. To overcome dephasing and prevent the electrons from outrunning the wakefield, spatiotemporal pulse shaping can be used to propagate the laser intensity at the speed of light in the plasma over long distances without the need for guiding structures. An axiparabola enables spatiotemporal control by focusing light rays at different nearfield radial locations to different farfield axial locations, while maintaining a small spot size over distances greater than a Rayleigh range. To control the time at which each radius comes to its corresponding focus, a radial group delay is introduced to the shape of the pulse. Two methods to achieve this are compared: (1) a reflective, radially stepped echelon optic and (2) two specially shaped glasses. Both techniques inherently introduce kvector spread and thereby farfield pulse broadening. The physical origins and implications of these chromatic aberrations are compared and discussed using scalar diffraction theory and raytrace modeling. 

PP11.00118: High Density Laser Wakefield Acceleration Ernesto BarrazaValdez, Bradley S Nicks, Toshiki Tajima High density Laser Wakefield Acceleration (HDLWFA) is studied in simulation using both beatwave and subcycle lasers near the critical density (ref.1). In contrast to the underdense regime of LWFA, the maximum electron energy in HDLWFA is shown to increase as the electron density increases toward the critical density. Energy transfer from the laser to the electrons obeys the same trend, indicating an increase in energy efficiency in the high density regime. These phenomena are potentially relevant for applications of HDLWFA in the fields of medical radiation therapy, sterilization, and microscopy. 

PP11.00119: Optimization of the performance of a laserplasma accelerator realized with an opticalfieldionized plasma channel Carlo Benedetti, Davide Terzani, Stepan S Bulanov, Anthony J Gonsalves, Marlene Turner, Carl B Schroeder, Eric Esarey Laserplasma accelerators can produce accelerating gradients on the order of tens to hundreds of GV/m, making them attractive as compact particle accelerators for radiation production or highenergy physics applications. Achieving large energy gains requires operating at low plasma densities and guiding the tightly focused laser driver over distances much longer than its characteristic diffraction length. Recently, several schemes for the production of meterscale plasma waveguides using opticalfieldionization (OFI) techniques have been proposed [1,2]. Compared to waveguides realized with a discharge capillary, which are characterized by a parabolic transverse density profile that extends over distances much larger than the laser spot size, and has ideal guiding properties, the density profile in OFI waveguides has a finite extension with a more complex structure made of a core and a finitewidth cladding. In this work we investigate, analytically and numerically, the propagation of a laser pulse in an OFI waveguide. We study the optimal matching condition that minimizes highorder modes beating, and we evaluate the laser depletion length and group velocity. Finally, the production of multiGeV electron beams using this type of waveguide is presented. 

PP11.00120: Towards a compact laser wakefield accelerator at kilohertz repetition rate Hao Ding, Anthony J Gonsalves, Tong Zhou, Liona FanChiang, Remi Lehe, Jeroen van Tilborg, Cameron R Geddes, James E Clayton, Eric Esarey Laser wakefield acceleration is a rapidly developing technology to make electron accelerators more compact and cost effective including for ultrafast particle and photon sources. To further mature this technology for demanding applications such as tabletop freeelectron lasers or inverse Compton scattering, stability, beam quality, and repetition rate need to be improved. The frequency of mechanical instabilities typically falls below 200 Hz. Operating a wakefield accelerator at a kHz rate thus not only offers the flux needed for applications, but also the opportunity for active feedback to stabilize the accelerator and to enable techniques that will improve beam quality. This poster reports on the recent progress on commissioning a kHz laser wakefield accelerator at Lawrence Berkeley National Laboratory. We have established a platform with a singlebox commercial offtheshelf kHz laser frontend, spectrally broadened in a hollowcore fiber, and subsequently compressed with chirped mirrors to sub5 fs pulse duration. Focusing these fewcycle pulses to relativistic peak intensities allows for resonant excitation of plasma waves, in which electrons can be accelerated to several MeVs of energy within a submillimeter scale plasma. 

PP11.00121: Controllable Interactions between CoPropagating Laser Pulses in Underdense Plasmas Nicholas Ernst, Yong Ma, Alec G.R. Thomas, Karl M Krushelnick In recent years, the generation of ultrarelativistic electron (e^{}) beams via LaserWakefield Acceleration (LWFA) has become routine  though control of beam properties and dynamics remains problematic. With the evolution of laser technology into petawatt (10^{15 }W) regimes, a unique opportunity arises to distribute power across multiple, copropagating beamlets. Previous work of these experimental situations has demonstrated a large parameter space with complex interaction of induced wakefields and e^{} beams. These mechanics are highly dependent on tractable parameters such as the number of beamlets, beamlet displacement, temporal delay, carrier frequency, etc. Here, we establish an overview of several multibeam effects and present how systematic investigation of beamlet parameters can lead to enhanced control over e^{} beam generation. ParticleinCell (PIC) simulations support this concept, where a lowintensity, satellite beamlet copropagating alongside a primary driver pulse can induce e^{} injection when operating in plasma densities below the selfinjection threshold of the driver. Promising results show control over injected bunch charge and duration. Supporting experiments for these multibeam setups are further outlined for future work. 

PP11.00122: Twocolor ionization injection in a multipulsedriven laserplasma accelerator Liona FanChiang, Anthony J Gonsalves, HaiEn Tsai, Tobias M Ostermayr, Davide Terzani, Carlo Benedetti, Carl B Schroeder, Samuel Barber, Jeroen van Tilborg, Curtis Berger, Cameron R Geddes, Eric Esarey Laser plasma accelerators (LPAs) have promise to be the next generation accelerator for a number of basic science, industry, security and medical applications. Not only are these accelerators extremely compact, they offer a path to orders of magnitude increase in brightness compared to conventional accelerators, enabled by low emittance and high peak current. High peak current naturally comes from the femtosecond length of LPA electron beams, but beam emittance needs to be improved. One proposal to achieve ultralow emittance from a laser plasma accelerator is an all optical, two laser configuration, where one is used to generate a plasma wake, while another, with low ponderomotive potential but high electric field, locally ionizes inner shell electrons, injecting them in the accelerating phase of the wake [1]. Since the electric field is proportional to the ponderomotive potential and inversely proportional to the wavelength, a high electric field with low ponderomotive potential can be achieved by using a smaller wavelength. We will present plans at BELLA Center to demonstrate this technique using pulses from a Ti:Sapphire laser system to generate the wake, and a third harmonic pulse to locally inject electrons into the wake. 

PP11.00123: Simulation Studies of Electron Injection in CO2 Laserdriven Selfmodulated and Blowout Regimes Roman V Samulyak, Prabhat Kumar, Aiqi Cheng, Rotem Kupfer, Michael C Downer, Vladimir N Litvinenko, Navid VafaeiNajafabadi, Irina Petrushina, Rafal Zgadzaj The process of electron selfinjection into CO2 laser driven plasma wakes in the parameter space ranging from the selfmodulated to the blowout regimes has been studied using 3D ParticleinCell simulations with code SPACE. In SMLWFA regime, selfinjection arises with the wave breaking which occurs at a field strength that is significantly below the 1D wavebreaking threshold. This process intensifies at higher laser power and plasma density and is suppressed at low plasma densities, below 1 ×10^{17} cm^{3}. In the blowout regime, the selfinjection was not observed under simulation conditions. The twocolor injection process driven in highZ material plasmas by a highintensity Tisapphire laser with low normalized vector potential a_{0}, followed up by a lowerintensity, higha_{0} CO2 laser pulse has also been investigated and compared with the recent experiments at Brookhaven National Laboratory.. 
Not Participating 
PP11.00124: Plasmaacceleratorbased linear beam cooling system Carl B Schroeder, Carlo Benedetti, Stepan S Bulanov, Davide Terzani, Eric Esarey, Cameron R Geddes Plasmabased accelerators enable compact acceleration of beams to high energy and are considered a promising technology for future colliders. Conventional colliders require damping rings to generate the required beam emittance for highenergy physics applications. We propose and discuss a plasmabased linear damping system that allows cooling of ultrashort bunches compatible with plasmabased accelerators, removing the need for bunch compression. Ultrahigh plasma accelerating gradients allow for linear damping systems, and there is a tradeoff between system length and the achievable emittance reduction. Final asymptotic beam emittance is shown to be independent of beam energy. 

PP11.00125: Recent Developments in QuickPIC Open Source Weiming An, Weiyu Meng, Rong Tang, Kaiguo Huang, Yueran Tian, Yueluo Wang, Viktor K Decyk, Yujian Zhao, Fei Li, Lance Hildebrand, Qianqian Su, Warren B Mori As a 3D parallel quasistatic PIC code, QuickPIC has been an open source code on Github since 2017 [1]. QuickPIC has been widely used for efficiently modeling the plasma based accelerator problems. Recently, we have merged the field ionization module from the old version of QuickPIC with the open source QuickPIC. We also implement a new field ionization module based on the particle ionization instead of the mesh ionization. The latter one can include mobile ions when simulating the field ionized plasma. We also implement a the new diagnostic for calculating the beam’s betatron radiation in QuickPIC. The beam and plasma particle tracking and the automatic memory reallocation for buffer overflows will be introduced. In addition, we will present the progress on merging the mesh refinement algorithm and the GPU algorithm into the open source code. 

PP11.00126: Prospects for muon acceleration in plasma based accelerators Chiara Badiali, Jorge Vieira, Chandrashekhar Joshi Muon colliders are being currently considered as a next step for HEP. Since muons, as electrons, are fundamental particles, their full energy is available in collisions, in contrast to protons. Nonetheless, the finite mean lifetime of muons (2.2 μs at rest) means that the muons must be collected, cooled, and rapidly accelerated before a significant number of them decay. Plasma accelerators could in principle accelerate muons to relativistic energies much faster than traditional RF accelerators, thereby greatly reducing the loss of muons. 

PP11.00127: Simulation and Measurement of Helicon Waves at the Madison AWAKE Prototype Marcel D Granetzny, Barret Elward, Michael Zepp, Oliver Schmitz The AWAKE project at CERN opens up the frontier of next generation electron colliders using beamplasma wakefield acceleration. Acceleration gradients exceeding 1 GV/m have been demonstrated using a laserionized plasma. However a full scale accelerator will need a reliable, high density plasma source that scales to kilometer lengths with a high degree of axial density uniformity. The Madison AWAKE Prototype (MAP) is utilizing 30 kW of RF power to generate a helicon plasma with expected densities reaching 10^{20} m^{3} in a multiantenna setup. In order to optimize the density profile an understanding of wave propagation and power deposition is essential. To this end we have developed a finite element model in Comsol that solves for the quasi 3D wavefields for a given temperature, density and neutral distribution. To do so the antenna currents are decomposed into discrete azimuthal modes and the 3D fields are reconstructed from solutions of the 2D axisymmetric problem for each dominant mode. We present simulation results for measured plasma profiles that reproduce experimentally seen effects such as m=1 mode dominance and asymmetric power deposition. Further we show the current development status of wavefield measuring diagnostics aiming to verify the simulations. 

PP11.00128: Current Filamentation Instabilities of Proton Beams in Proton Driven Wakefield Accelerators Erwin Walter, Martin S Weidl, John P Farmer, Patric Muggli, Frank Jenko Plasma wakefield accelerators can generate electricfield gradients magnitudes larger than conventional accelerators. Using this technology, particlephysics experiments could be performed in much more compact devices. 

PP11.00129: High Time Resolution Axial Particle Balance in AWAKE Helicon Plasmas Michael Zepp, Birger Buttenschön, Barret Elward, Marcel D Granetzny, Jonathan Green, Oliver Schmitz, Alban Sublet A high density plasma (n_{e }≈ 10^{21} m^{3}) is needed to achieve Wakefield acceleration of electrons from an axial electric field in the GV/m range. It has been shown that sufficient densities are achievable in helicon plasmas during 5 ms pulses. It is still necessary to determine the axial density homogeneity, which is required to stay within 0.25% for Wakefield acceleration. An argon laser induced florescence technique to measure the axial density homogeneity and derive a particle balance on submillisecond timescales is being developed. Acoustooptic modulators have been successfully tested for frequency shifting to avoid the need for the typical frequency scan. Utilizing these frequency shifters, the new diagnostic will be capable of acquiring 12point density data after only one 5ms pulse, and ion velocity data and an axial particle balance after only two 5ms pulses. The particle balance will indicate if additional fueling is necessary to correct the axial density homogeneity. 

PP11.00130: Electron beam probing: essentials for the direct measurements of plasma wakefields Irina Petrushina, Apurva Gaikwad, Rafal Zgadzaj, Igor Pogorelsky, Marcus Babzien, Mikhail Fedurin, Rotem Kupfer, Karl Kusche, Mikhail Polyanskiy, Mark A Palmer, Roman V Samulyak, Chaojie Zhang, Warren B Mori, Michael C Downer, Chandrashekhar Joshi, Vladimir N Litvinenko, Navid VafaeiNajafabadi Direct experimental characterization of plasma wakefields is highly valuable for understanding the wake evolution in various regimes as well as the injection mechanisms. While optical methods of wakefield probing have been extensively studied and have successfully demonstrated the ability to capture shadowgraphs of the plasma wakes, such methods directly rely on the change of the index of refraction and are not applicable for diagnostic of lowdensity plasmas. Instead, an ultrashort relativistic electron beam can be used to visualize the plasma wakefields: when passing through a wake, an electron beam density is modulated by the electric fields of the wake. Once imaged, this density modulation provides an insightful snapshot of the wake structure allowing for the direct characterization of the wakefield. We present the essential considerations necessary to implement the direct wakefield visualization technique using an electron beam as a probe. The discussion is focused on the electron probing experiment implemented at the Accelerator Test Facility (ATF) of Brookhaven National Laboratory (BNL). 

PP11.00131: MACHINE LEARNING MINICONFERENCE


PP11.00132: A DataDriven Approach Towards NIF Neutron TimeOfFlight Diagnostics Using Machine Learning and Bayesian Inference SuAnn Chong, Dave Schlossberg, Jim A Gaffney, Luc Peterson, Kelli D Humbird The neutron timeofflight (nToF) diagnostic is used to diagnose implosion dynamics during inertial confinement fusion (ICF) experiments. The primary goal of the nToF diagnostic is to extract important fusion quantities such as neutron yield, ion temperature and downscatter ratio from neutron spectra to study the dynamics of thermonuclear fusion. In this work, we present a datadriven approach as an alternative to a physicsdriven approach for nToF diagnostics at the National Ignition Facility (NIF). Instead of deriving point estimates of fusion quantities, our approach offers an approximation of the posterior distribution of the fusion quantities using a Markov chain Monte Carlo (MCMC) method. In the event of insufficient ICF experimental data, simulation outputs are needed. However, running complex simulations are computationally expensive. Hence, to speed up the data generation process, we trained a Deep Jointly Informed Neural Network (DJINN) that serves as a surrogate model to generate simulation outputs. The results of our approach and a comparison with the nToF physicsdriven approach are presented in this work. Our approach is an attractive alternative to understand complex systems especially when physical models lack a complete description of the system of interest. 

PP11.00133: Classifying Alfvén eigenmodes using deep learning and CO_{2} interferometer data at DIIID Alvin V Garcia, Azarakhsh Jalalvand, William W Heidbrink, Egemen Kolemen, Michael A Van Zeeland Deep learning models are trained using CO_{2} interferometer data and human labels to detect the presence of Alfvén eigenmodes (AE) in 1112 discharges at DIIID. The goal of this project is to demonstrate the potential of using deep learning models to accurately detect AE modes observed in tokamak plasmas. Resonant fast ions can drive AE modes unstable and degrade the plasma performance or energy confinement. Mode activity can be detected in the crosspower spectra of the CO_{2} interferometer diagnostic. This task is time intensive and requires extensive domain knowledge. However, implementing machine learning models into the analysis routine would accelerate the process and increase the frequency of accurate predictions. Recent work produced a database of the occurrence of EAE, TAE, RSAE, BAE, LFM, and EGAM activity [Heidbrink, et al., NF ‘20] that is suitable for machine learning analysis. This poster discusses the application of deep learning algorithms to classify AE modes in the CO2 interferometer dataset. Preliminary results show good predictions of TAE and RSAE. Supported by the U.S. Department of Energy under DESC0020337, DEFC0204ER54698 and National Science Foundation under 1633631. 

PP11.00134: Development of Machine Learning Informed Design Optimization for Double Shell Capsule Graded Density Inner Layer Targets Nomita Vazirani, David Stark, Paul A Bradley, Michael J Grosskopf, Eric N Loomis, Brian M Haines, Scott England, Wayne Scales Advances in machine learning (ML) have the ability to reduce design costs and enhance the design process in HED experiments. ML can use a minimal number of expensive simulations to search the design space efficiently for optimal designs of ICF experiments. This work focuses on optimizing graded density inner shells of indirectly driven double shell targets (Phys. Plasmas, 26, 052702 (2019)), while reducing hydrodynamic instability and maintaining high yield. Graded layer inner shell targets have a parameter space that is too large to fully map out, which is why efficient exploration of the design space is not only beneficial but also necessary. ML methods use predictive physics simulations to identify graded layer designs with high predicted performance as well as novel designs with high uncertainty in performance that may hold unexpected promise. We apply Bayesian optimization to the design optimization of double shell graded inner shell targets using physics models of varying fidelities. By applying ML tools to design simulations, we aim to optimize target geometry to mitigate hydrodynamic instability and improve yield. We present our progress on this new ML informed design tool for future application to NIF double shell experiments with graded layer inner shell targets. 

PP11.00135: Physicsinformed deep learning of dynamic 3D experiments Zhehui Wang, Bradley T Wolfe, J S BenBenjamin, Chris S Campbell, Dana M Dattelbaum, Kamel Fezzaa, John E Foster, Zhizhong Han, Hanna Kim, John L Kline, Yao E Kovach, Nga T NgyuyenFotiadas, Christopher Roper, Yancey H Sechrest, David Staack, Xin Tang, Miles T TengLevy Successful applications of deep learning for data analysis such as classification and nonlinear regression are sometimes perceived as a ‘blackbox’ magic. This is not a satisfying situation for experimental physics, where interpretation of observations through the framework of fundamental and extended physics is essential. Meanwhile, continuing advances in hardware and different imaging modalities such as visible light, Xray and neutron, make image data increasingly accessible and as a result, machine learning appears to be unavoidable for automated information extraction from the oceans of data [1, 2]. Recent progress in data science indicates that physics and other knowledge such as geometry and material properties may be imbedded in physicsinformed or physicsconstrained deep neutral networks through a variety of means such as synthetic data generation, experimental data augmentation, automated uncertainty quantification, transfer learning, regularization by differential equations or scaling laws. Here we review and highlight recent progress in combining image data, deep learning, and physics models for dynamic experiments in a diverse set of 3D geometries: exploding wire experiments, ultrafast multiphase processes in liquids, plasmaliquid interfaces, shock compression of materials, and laser/Xray compression of materials. Our work aims at a holistic approach to a.) highspeed imaging, data science consistent with physics and humanlevel understanding and b.) experimental physics and material science involving dynamic 3D or 4D scenes. This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility, operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DEAC0206CH11357. The work is supported in part by LANL C2, LANL ICF, LANL LDRD, and NSFMSGI programs. 

PP11.00136: Physically Consistent Neural Network Models for Equations of State Luc Peterson, Katherine Mentzer Material equations of state provide key information to multiphysics simulation codes, like those used to model inertial confinement fusion and high energy density experiments. For complicated realworld materials, discrete tables that combine theory, simulation and experimental data are the current state of practice. However, when interpolating between tabular data points, physical inconsistencies can arise merely due to the interpolation process. In regions of discontinuities, such as around phase transitions, spurious oscillations in the interpolator can become particularly problematic and can even break physical consistency. In this work, we explore the use of neural networks as an efficient global interpolator for tabular equation of state data and develop model architectures that can not only recreate the original tabular data to within a few percent, but also respect phase boundary transitions without spurious oscillation. As a proof of concept, we apply the methodology to materials of interest to nuclear fusion simulations, such as deuteriumtritium mixtures. 

PP11.00137: Plasma Kink Classification Using Deep Learning Miles T TengLevy, Bradley Wolfe, Yi Zhou, Ryan S Marshall, Paul M Bellan, Zhehui Wang Contacts: mtenglevy@lanl.gov, bwolfe@lanl.gov 
Not Participating 
PP11.00138: Machine learning models for realtime, high bandwidth inference of ELM events and confinement regime with 2D BES at DIIID David R Smith, Prannav Arora, Lakshya Malhotra, George McKee, Zheng Yan, Mark D Boyer, Ryan Coffee, Azarakhsh Jalalvand, Egemen Kolemen Multichannel fluctuation diagnostics capture the spatial patterns of highbandwidth plasma dynamics. Here, we report on an effort to develop machine learning (ML) models for the realtime identification of edgelocalizedmode (ELM) events and the turbulence properties of confinement regimes using the 2D Beam Emission Spectroscopy (BES) system at DIIID. The "edge ML" models will be deployed on a highthroughput FPGA accelerator for integration in the realtime plasma control system (PCS). The models will generate reduced signals that correspond to ELM activity and turbulence dynamics, and the realtime PCS will learn to avoid ELM regimes and to steer the plasma towards and maintain advanced confinement regimes such as the wide pedestal QHmode. The 2D BES system captures plasma density perturbations imprinted in neutral beam emission at a 1 MHz frame rate. The edge ML models will analyze about 10 ms histories from the BES data stream to assess ELM and turbulence activity. Preliminary results for classifying active ELM events give a ROCAUC score of about 0.98 for validation data. We also explore different neural network architectures such as autoencoders to compress spatiotemporal information in lowdimension feature space for multiple classification and prediction tasks. 

PP11.00139: Structure identification, tracking and velocimetry in gaspuff imaging measurements Mate Lampert, Ahmed Diallo Field aligned structures play a key role in plasma particle and heat transport, thus, their investigation is of great interest. Intermittent structures such as blobs and transient structures such as ELM filaments have been observed routinely on NSTX by the means of gaspuff imaging (GPI). GPI provides fast temporal (2.5us) and high spatial resolution (1cm) measurement of these structures. Thus far the analysis techniques applied to GPI data has not expoited the high resolution of the measurement to its fullest. 
Not Participating 
PP11.00140: Machine learning autoencoder models for compressing and reconstructing 2D BES data for realtime classification tasks. Prannav Arora, Lakshya Malhotra, George McKee, David R Smith, Zheng Yan, Mark D Boyer, Ryan Coffee, Azarakhsh Jalalvand, Egemen Kolemen Multichannel fluctuation diagnostics capture the plasma dynamics. Here, we report on an effort to develop machine learning (ML) models for the realtime identification of edgelocalizedmode (ELM) events and the turbulence properties of confinement regimes using the 2D Beam Emission Spectroscopy (BES) system at DIIID. The "edge ML" models will be deployed on a highthroughput FPGA accelerator for integration in the realtime plasma control system (PCS). The models will generate reduced signals that correspond to ELM activity and turbulence dynamics, and the realtime PCS will be trained to avoid ELM regimes and to maintain advanced confinement regimes such as the wide pedestal QHmode. The 2D BES system captures density perturbations imprinted in neutral beam emission at a 1 MHz frame rate. Here, we report on autoencoder neural networks to compress the spatialtemporal information in a lowdimension space. Using such an autoencoder, we plan to compress BES data for ELM classification and other classification tasks. Currently, we experiment with the number of hidden layers and network architecture to maximize the capability of the network to compress and reconstruct 2D BES data as measured by a mean squared error loss function. 

PP11.00141: Analysis of MHD Modes in Predicting ELM Onset using Machine Learning Techniques Jeffrey Zimmerman, Ahmed Diallo, Christopher Battista Detection and prevention of edge localized modes (ELMs), which are destructive disturbances in the edge regions of Hmode toroidal plasmas, pose a significant roadblock to the development of stable, efficient fusion reactors. This problem is compounded by the fact that while models have been proposed to explain mechanisms behind them, ELMs are widely considered to be beyond the stateoftheart predictive modeling, resulting in limited theoretical understanding of ELM onset to date. Machine learning (ML) has demonstrated effectiveness on other difficult plasma physics problems, such as predicting turbulent fluxes and estimating thermodynamic profiles from sensing data. Here we apply two machine learning models to predict ELM onset directly from Mirnov coil sensor data taken from a General Atomics DIIID tokamak. The ML models support a prediction horizon long enough for realistic realtime ELM detection and mitigation. The models (multivariate polynomial regression and multilayer perceptron neural network) perform better than chance, with mean R^{2} values of 0.39 and 0.45, respectively. In addition, we extract analytical representations from the polynomial regression model, yielding relationships between ELM occurrence and the presence of resonance modes in the magnetic flux density of fusion reactors. 

PP11.00142: Resistive Wall Mode stability forecasting in NSTX through Balanced Random Forests and counterfactual explanations Andrea Piccione, John Berkery, Steven A Sabbagh, Yiannis Andreopoulos Recent progress in the Disruption Event Characterization and Forecasting (DECAF) framework [1] has shown that Machine Learning (ML) guided by physics theory [2] can be easily implemented as a supporting tool for fast computations of ideal stability properties of spherical tokamak plasmas. In order to extend that idea, a customized Random Forest (RF) classifier that takes into account imbalances in the training data is hereby employed to predict Resistive Wall Mode (RWM) stability for a set of high beta discharges from the NSTX spherical tokamak. More specifically, with this approach each tree in the forest is trained on bootstrap samples that are balanced via a userdefined over/undersampler. The proposed approach outperforms classical costsensitive methods for the problem at hand, in particular when used in conjunction with a random undersampler, while also resulting in a threefold reduction in the training time. In order to further understand the model's decisions, a diverse set of counterfactual explanations based 

PP11.00143: Datadriven Techniques for Time Domain Decomposition of Plasma Physics Simulation Sebastian De Pascuale, Kenneth Allen, David L Green, Jeremy D Lore We present the analysis of datadriven techniques for dimensionality reduction of plasma physics simulation. Surrogate models derived from limited output of such codes offer an approach to lower the computational burden of prohibitive calculations in high spatial or temporal regimes. We target the application of these models towards representation of time domain features in plasma edge simulations of the tokamak scrapeoff layer boundary obtained from SOLPSITER. Starting from lowrank matrix approximation, we showcase a principled parameter tuning for selection of sampled simulation snapshots, construction of operators for time advance, and estimation of reduced model accuracy. We consider two use cases for the proposed techniques: extraction of bestfit linear dynamics from nonlinear data and identification of CUR decompositions as an efficient alternative to the SVD. In the former case, dynamic mode decomposition provides a method to segregate timescales and extrapolate forward stable features. In the latter case, skeleton decomposition provides an interpretable representation of matrix data optimized over physical coordinates and recorded timesteps. We discuss the implementation of these techniques for the acceleration and compression of plasma physics simulation. 

PP11.00144: Quantifying Uncertainty in EquationofState Models with Thermodynamically Constrained Machine Learning Jim A Gaffney, Suzanne J Ali, Lin H Yang Equationofstate (EOS) models provide an interesting challenge for uncertainty quantification and machine learning since they must be trained on very sparse data and then extrapolated over huge regions of the input parameter space. They are also subject to strong physical constraints; even small deviations from socalled thermodynamic consistency results in the failure of downstream tasks, for example radiationhydrodynamics simulations. On the other hand, the constraints are known analytically and so can be enforced by choosing a suitable form for the EOS. 

PP11.00145: New Neural Networks for Plasma Profile Prediction Andrew Rothstein, Egemen Kolemen As machine learning has become more accessible, it has become an adaptable, powerful tool that can reduce computational costs for modeling complex systems. Since modeling the plasma in a tokamak reactor is both impractical and unreliable, it is more advantageous to use previous reactor data and current diagnostics to predict how the plasma will evolve in the future. As the accuracy and speed of these models develop, they have the potential to become the core of tokamak reactor control systems. We sample different types of neural networks to predict plasma conditions inside tokamaks to find models that are better suited for active control of tokamaks. To analyze the utility of each type of model, we compare the computational costs of training, the accuracy of the models, and the speed for a new prediction to be made. 

PP11.00146: Timedependent analysis of edge plasma turbulence via deep learning from partial observations Abhilash Mathews, Jerry W Hughes, Manaure Francisquez, James L Terry, Noah R Mandell, Seung Gyou Baek, Adam Q Kuang, David R Hatch We demonstrate that a physicsinformed multinetwork deep learning architecture constrained by partial differential equations can accurately learn turbulent fields consistent with driftreduced Braginskii theory from just partial observations of electron pressure in contrast with conventional analytic equilibrium methods. This framework further enables the first ever direct quantitative comparisons of electron pressure and electric field fluctuations in nonlinear global electromagnetic gyrokinetic simulations and electrostatic twofluid theory. Accordingly, we quantitatively explore the concomitant response that exists between the fluctuating electron pressure and electric potential which constitutes one of the key relationships demarcating a plasma turbulence model. The methods outlined can be readily adapted to the study of magnetized quasineutral plasmas in advanced geometries and presents broad implications for the validation of reduced plasma turbulence models in experimental and astrophysical settings. In particular, applications of our deep learning framework to tokamak diagnostics for timedependent analysis and interpretation of edge turbulent fluctuations measured by gas puff imaging will be considered. 

PP11.00147: Particleincell simulations of laserplasma instabilities: creation of a dataset for training a deeplearningbased LPI surrogate for ICF Alessandro Ruocco, Robbie H Scott, William Trickey, Sam M Vinko, Muhammad F Kasim For the success of inertial confinement fusion (ICF) [1], it is crucial to understand some of laserplasma instabilities (LPIs) [2] effects, particularly i) scattered light, and ii) features of suprathermal electrons. However, LPI complexity obscures experimental characterisation and mathematical understanding of those processes. In addition, the LPI kinetic nature makes them incompatible to be fully included in ICF hydrodynamics codes. Then, in hydrocodes, LPI effects are simplified using scaling laws based on a few experimental and/or simulations results due to limited resources [3], constraining the hydrocodes accuracy, especially at ignition conditions. 

PP11.00148: Optimisation of highintensity lasersolid interactions using gaussian process regression. Charlotte A Palmer, Matthew J. V Streeter, Brendan Loughran, Hamad Ahmed, Sam Astbury, Marco Borghesi, Nicolas Bourgeois, Chandra Breanne Curry, Stephen J Dann, Nicholas P Dover, Tom Dzelzainis, Oliver Ettlinger, Maxence Gauthier, Lorenzo Giuffrida, Griffin Glenn, Siegfried Glenzer, Ross Gray, James Green, George Hicks, Cormac Hyland, Valeriia Istokskaia, Martin King, Daniele Margarone, Orla McCusker, Paul McKenna, Zulfikar Najmudin, Claudia Parisuana, Peter Parsons, Chris Spindloe, Dan R Symes, Franziska Treffert, Nuo Xu, Alec G.R. Thomas Laserdriven energetic proton accelerators have the potential to provide compact sources of MeV energy, low emittance, subpicosecond duration proton beams for a variety of applications. The primary impediment to their wider adoption is the challenge of shottoshot reproducibility and tuning of the parameters to optimize desirable proton beam qualities in a multidimensional parameter space. Recent developments in laser technology and control systems, making available multiHz delivery of jouleclass, relativisticallyintense laser pulses with automated control, combined with online diagnostics have enabled the automated scanning of parameters space, quantification of uncertainty and use of feedback loops for optimization of desirable outputs (e.g. proton beam maximum energy). Bayesian optimization has already demonstrated impressive gain in xray generation when used in conjunction with a laser wakefield accelerator [1]. Here, we discuss the preliminary results from experiments expanding this tool to laserdriven proton acceleration and challenges facing the adaption of this gaussianprocessregressionbased Bayesian optimizer to the sharply varying parameter space of lasersolid interactions. 

PP11.00149: Machine learning directed TGLF saturation rule development Sterling P Smith, Tom F Neiser, Adam Eubanks, Orso Meneghini, Gary M Staebler, Jeff Candy To validate the TGLF model in general, and the saturation rules in particular, we have built a curated database of experimental profiles and power balance analyses of DIIID plasmas using an automated workflow within OMFIT. After successful validation, we applied machine learning tools to this dataset to help direct future model development. As a first step, we applied model optimization tools to the free parameters in SAT0 and SAT1, which have previously been calibrated against a database of gyrokinetic simulations with GYRO. As a second step, machine learning tools were employed to direct SAT1 model development. The SAT1 intensity spectrum was multiplied with a `correction factor' of the form $(a/k^c)/\exp(b/k)$, where `k' is the binormal wavenumber, `b' is some constant and `a' and `c' are outputs of a neural network. Using system identification tools, we found that `a' and `c' are best described by a power law of plasma parameters that typically affect TEM turbulence. As third step, we used the above workflow independent of existing saturation rules to find new (or confirm existing) avenues of saturation rule development. A hypothetical saturation rule is constructed and all free parameters are calibrated against a database of gyrokinetic simulations. Comparison to experiment is then used to validate the development workflow of existing saturation rules. 

PP11.00150: A ModelBased Reinforcement Learning Approach for Beta Control Ian Char, Youngseog Chung, Mark D Boyer, Egemen Kolemen, Jeff Schneider The goal in Reinforcement Learning (RL) is to learn from data a controller that can take optimal actions in an environment. Recently, RL has produced astounding results in a variety of applications, including the game of Go, video games, and robotic control. In this work, we explore whether recent strides in this field can be applied to the plasma physics setting to derive a highperforming controller for Beta target tracking. 

PP11.00151: Normalizing flows for likelihoodfree inference with fusion simulations Michael Churchill, Chirag Furia Fluid based scrapeoff layer transport codes such as UEDGE are heavily utilized in tokamak analysis and design, but typically require userspecified anomalous transport coefficients to match experiment. Determining uniqueness of these parameters and the uncertainties in them to match experiments can provide valuable insights to fusion scientists. We leverage recent work in the area of likelihoodfree inference ("simulationbased inference") to train a neural network which enables accurate statistical inference of the anomalous transport coefficients given experimental plasma profile input. UEDGE is treated as a blackbox simulator, and run multiple times with anomalous transport coefficients sampled from priors, and the neural network is trained on these simulations to emulate the posterior. The neural network is trained as a normalizing flow model for density estimation, allowing it to accurately represent complicated, highdimensional distribution functions. We discuss important implementation details such as the use of summary statistics and the number of simulations needed for good results. We also discuss the future possibilities for use of amortized models which train on a wide range of simulations and enable fast statistical inference for results during experiments. 

PP11.00152: Transfer Learning for the Reproduction of HighFidelity Opacity Spectra Michael D Vander Wal, Kelli D Humbird, Ryan G McClarren Simulations are an important part of the inertial confinement fusion (ICF) experiment design; however, these simulations have high computational costs. One of the primary contributors to the cost is the calculation of the nonlocal thermal equilibrium opacities which can consume anywhere from 10% to 90%+ of the computational time. Previous studies (Deep Learning for NLTE Spectral Opacities, Gilles Kluth, 2020) demonstrate that 7x speeds up of hohlraum simulations are achievable by replacing the atomic physics calculation with a trained neural network emulator trained on several hundred thousand Cretin (Kluth) calculations. This work trained on a database of moderate fidelity DCA data for a single element. 

PP11.00153: Single Gaussian Process Method for Arbitrary Tokamak Regimes Jarrod Leddy, Sandeep Madireddy, Eric C Howell, Scott E Kruger Gaussian Process Regression (GPR) is a Bayesian method for inferring profiles based on input data. The technique is increasing in popularity in the fusion community due to its many advantages over traditional fitting techniques including intrinsic uncertainty quantification and robustness to overfitting. Most fusion researchers to date have utilized a different GPR kernel for each tokamak regime. This requires a Machine Learning (or simpler) method to first predict the regime, choose the right kernel for that regime, and then use that kernel. The disadvantage of this method is that it requires an additional step, and it is unclear how well it will behave if the plasma enters a new, unexpected regime. We summarize our work developing a general kernel for all regimes (including radiallyvarying hyperparameters), utilizing heavytailed likelihood distributions to automatically handle data outliers, and using GPflow for full Bayesian inference via Markov chain Monte Carlo to sample hyperparameter distributions. We present a single GPR method that is robust across many different tokamak regimes and a wide range of data inputs and quality. 

PP11.00154: Spatiallylocalized Alfvén eigenmode classification using convolutional neural networks Alan Kaptanoglu, Azarakhsh Jalalvand, Alvin V Garcia, Andrew O Nelson, Joseph A Abbate, Geert Verdoolaege, Steven L Brunton, William W Heidbrink, Egemen Kolemen We use an expertlabeled database of DIIID discharges to classify five types of Alfvén eigenmodes (AEs) with convolutional neural networks, opening up the possibility of deeplearningenhanced realtime control of this important class of plasma dynamics. Each DIIID discharge in the database consists of forty radiallylocalized electron cyclotron emission (ECE) measurements, sampled at 500 kHz for the first 2 seconds of the discharge. The model attempts to predict when each AE type occurs in a validation dataset, and discriminates between the five types of AE activity. This strategy performs strongly at spatiotemporally localized prediction and classification of Alfven eigenmodes (approximate average true positive rates of 80% and false positive rates of 2%), indicating future promise for more sophisticated spatiotemporal models and incorporation into future realtime control strategies. 

PP11.00155: Adding image diagnostics to the prediction of future ICF experiments at NIF Bogdan Kustowski, Kelli D Humbird, Brian K Spears, Jim A Gaffney, Eugene Kur Recently, a new machine learning workflow has been developed to predict the outcome of the future indirect drive ICF experiments at the National Ignition Facility. Given the design parameters of an upcoming experiment, a 2D integrated hohlraumcapsule simulation is first run to make an initial prediction of multiple diagnostic measurements. Then, a machine learning model improves this prediction using a database of the past experiments. In this presentation, we extend this workflow to match not only scalar but also image data. We will discuss technical challenges triggered by including image data and compare the new, imageinformed predictions with the scalaronly predictions and with the real experimental data. We will investigate how matching more types of diagnostics helps break degeneracies in the model and reduce error bars. Ultimately, an improved model will be used in the exploration of new ICF designs. 

PP11.00156: Estimating the probability of a rare event using imprecise probabilities (IP)/ probability bounds analysis (PBA) Nelson M Hoffman If we must estimate the probability of a rare event, given little data and inadequate models, we are in deep water, to put it mildly. To make matters worse, in such instances, ordinary Bayesian inference relies heavily on modelparameter priors that are perhaps unrealistically precise. The method of imprecise probabilities (IP) or probability bounds analysis (PBA)^{1} is a way of enforcing greater humility regarding the true state of our prior knowledge, and thus the range of feasible priors. Picard and Vander Wiel^{2} describe an example based on highexplosive droptest data, where the detonation of a highexplosive sample on which a heavy weight is dropped is the (undesirable) rare event. We reproduce their example using the GPMSA^{3} Bayesian inference code, and show how probability estimates depend on user choices regarding the allowable range of priors, the model form, and the search design in modelparameter space. We go on to consider the case of inertialconfinement fusion, where fusion ignition is the (desirable) rare (thus far) event. 

PP11.00157: An unsupervised machinelearning checkpointrestart algorithm using Gaussian mixtures for particleincell simulations Guangye Chen, Luis Chacon, Truong Nguyen We propose an unsupervised machinelearning checkpointrestart (CR) lossy algorithm for particleincell (PIC) algorithms using Gaussian mixtures (GM). The algorithm fea tures a particle compression stage and a particle reconstruc tion stage, where a continuum particle distribution function is constructed and resampled, respectively. To guarantee fidelity of the CR process, we ensure the exact preservation of charge, momentum, and energy for both compression and reconstruction stages, everywhere on the mesh. We also ensure the preservation of Gauss’ law after particle reconstruction. As a result, the GM CR algorithm is shown to provide a clean, conservative restart capability while potentially affording orders of magnitude savings in input/output requirements. We demonstrate the algorithm using a recently developed exactly energy and chargeconserving PIC algorithm on physical problems of interest, with compression factors > 75 with no appreciable impact on the quality of the restarted dynamics. 

PP11.00158: Development of a Deep Neural Network Model for Spacecraft Charging Sayan Adhikari, Rupak Mukherjee, Sigvald Marholm, Wojciech J Miloch A Deep Neural Network (DNN) model using the kinetic ParticleinCell (PIC) simulation data for spacecraft charging has been developed. The model will be able to predict the potential structure around the spacecraft based on the system parameters of any given system with certain accuracy. DNN provides an excellent tool where we could leverage deep learning for all machine learning tasks and expect better performance with surplus data availability. The usability of DNN is unlimited if a user can train such a model with physicsbased parameters. In recent studies, it has shown promising outcomes in terms of accurate prediction of physical quantities[12]. In the present work, we used Tensorflow[3], a deep learning library, to classify the PIC simulation datasets. Using Tensorflow, we have compared the effects of multiple activation functions on classification results and developed a library for a case study. This work shows that integrating DNN into traditional computational methods might be the new beginning of developing nextgeneration modeling. 

PP11.00159: Learning Augmentation from Data: A Case Study in Scientific Diagnostic Simulations Sam A Jacobs, Tuan Tran, Derek Mariscal, Tim Moon, Blagoje Djordjevic, Michael Wyatt, Brian Essen, Tammy Ma Deep neural networks are a popular machine learning technique for analyzing and extracting useful information from data. However, neural networks are usually overparameterized and require large amounts of example training data to avoid overfitting and maximize generalizability. For image classification applications, data augmentation has shown great success as a means of increasing the effective size of datasets. However, automated data augmentation for scientific applications has not been thoroughly explored. Scientific datasets would especially benefit from data augmentation since simulations and experimental data can be prohibitively expensive. In this paper, we explore the idea of populationbased augmentation (PBA) to learn on scientific datasets. PBA involves multiple neural networks, each with different data augmentation schedules and policies, that are trained simultaneously and evaluated at intervals to find the best augmentation policy. We provide experimental results from an automated analysis of proton beam imaging energy spectrometer data (PROBIES). 

PP11.00160: Machine learning models for realtime, high bandwidth inference of ELM events and confinement regime with 2D BES at DIIID Lakshya Malhotra, Prannav Arora, George McKee, David R Smith, Zheng Yan, Mark D Boyer, Ryan Coffee, Azarakhsh Jalalvand, Egemen Kolemen Multichannel fluctuation diagnostics capture the spatial patterns of highbandwidth plasma dynamics. Here, we report on an effort to develop machine learning (ML) models for the realtime identification of edgelocalizedmode (ELM) events and the turbulence properties of confinement regimes using the 2D Beam Emission Spectroscopy (BES) system at DIIID. The "edge ML" models will be deployed on a highthroughput FPGA accelerator for integration in the realtime plasma control system (PCS). The models will generate reduced signals that correspond to ELM activity and turbulence dynamics, and the realtime PCS will learn to avoid ELM regimes and to steer the plasma towards and maintain advanced confinement regimes such as the wide pedestal QHmode. The 2D BES system captures plasma density perturbations imprinted in neutral beam emission at a 1 MHz frame rate. The edge ML models will analyze about 10 ms histories from the BES data stream to assess ELM and turbulence activity. Preliminary results for classifying active ELM events give a ROCAUC score of about 0.98 for validation data. We also explore different neural network architectures such as autoencoders to compress spatiotemporal information in lowdimension feature space for multiple classification and prediction tasks. 

PP11.00161: Inferring Lorentzcovariant reduced plasma models from fullykinetic simulations Madox McGraeMenge, Jacob R Pierce, Frederico Fluza, E. Paulo Alves Techniques from data science and machine learning are offering new approaches to developing theoretical and computational models of plasma dynamics directly from data. In particular, recent work has demonstrated the possibility of leveraging sparse regression techniques to recover interpretable plasma models [in the form of partial differential equations (PDEs)] from the data of fullykineticparticleincell simulations. However, to robustly apply this methodology to uncover new reduced models of poorly understood plasma dynamics, it is important to embed fundamental physical constraints and symmetries in the inference procedure. Here, we show that embedding these known physical symmetries through dataaugmentation is highly effective in improving the accuracy and robustness of the inferred PDEs. We specifically focus on enforcing Lorentz covariance of the inferred PDE models, which we achieve by Lorentzboosting the data into reference frames moving at random velocities. We demonstrate the benefits of this approach on the inference of the fundamental hierarchy of plasma models, from the kinetic Vlasov equation to the singlefluid plasma equations, from PIC simulation data. We show that using Lorentzaugmented data leads to 1) more accurate identification of model coefficients (up to three orders of magnitude more accurate compared with using original labframe data alone), and 2) the elimination of spurious unphysical PDE terms that do not satisfy Lorentz covariance. We further show that this data augmentation approach greatly relaxes the amount of original labframe data from expensive kinetic simulations needed for accurate and robust inference of reduced plasma PDE models. 

PP11.00162: Comparative Study of Neural Networks for 3D Reconstruction of 2D XRay Radiograph Bradley Wolfe, Zhizhong Han, Jonathan S BenBenjamin, John L Kline, Zhehui Wang Xray radiography acts as an important diagnostic for tracking the evolution of inertial confinement fusion targets and the growth of shell asymmetries. Traditionally, the analysis of the radiographs use information such as 2D contours to determine the modes of asymmetry. This method of analysis fails to account for 3D shell asymmetry which is present in the experiments. To account for various 3D features, we describe neural networks to reconstruct 3D models from experimental radiographs. We use these 3D models in order to determine 3D modes of asymmetry using various representations such as wavelets, Fourier modes and spherical harmonics. The evolution of 3D features of the ICF shells such as the joint between shell hemispheres has been characterized. We use multiple neural networks in order to compare the robustness and uncertainties of the results. 

PP11.00163: Using reinforcement learning to automate mesh management for HYDRA simulations Jay D Salmonson, Christopher K Yang, Christopher V Young, Joseph M Koning Multiphysics HYDRA simulations for inertial confinement fusion (ICF) experiments at the National Ignition Facility use mesh relaxation directives to manage the state of the arbitrary LagrangianEulerian (ALE) mesh and prevent entanglement. The difficulty of anticipating when and why mesh tangles occur and crash a simulation has historically required laborious manual intervention as well as a strategy to play it safe with overrelaxation at the expense of fidelity. An automated solution that can adapt and learn new situations would improve robustness and fidelity of simulations and save significant user time. To this end we have developed an unsupervised reinforcement learning method for managing ALE. It is made of two deep learning neural nets. The first is a convolutional neural net (CNN) that looks at patches of the mesh and is trained to predict how that mesh would evolve without intervention; it returns an image of the future of the mesh patch. The second net is a Soft ActorCritic (SAC) deep reinforcement learning algorithm that learns how to apply mesh relaxation thresholds that receive the highest reward, which is designed to improve the mesh quality the most with the least intervention. As the simulation runs, this system identifies problem areas in the mesh and applies a relaxation policy to improve the mesh condition with minimal intervention. 

PP11.00164: Machine learning for tokamak scenario optimization: combining accelerating physics models and empirical models Mark D Boyer, Josiah T Wai, Mitchell D Clement, Egemen Kolemen, Ian Char, Youngseog Chung, Willie Neiswanger, Jeff Schneider Betweenshots and realtime actuator trajectory planning will be critical to achieving high performance scenarios and reliable, disruptionfree operation in presentday tokamaks, ITER, and future fusion reactors. These tools require models that are both accurate enough to facilitate useful decision making and fast enough to enable optimization algorithms to meet betweenshots and realtime deadlines. While stateoftheart integrated modeling codes come close to the accuracy and completeness needed for these applications, they are too computationally intensive. To address this problem, a novel accelerated simulation capability has been developed by applying machine learning techniques to both empirical data and physicsbased simulations, enabling profile and equilibrium predictions at realtime relevant time scales. Coupled with numerical optimization schemes, the results provide a fast tool to design experiments or guide exploration of operating space. Use of AI/ML enables models to execute fast enough for use in realtime applications while maintaining high accuracy. The approach uses physics models for phenomena that are well described by models (e.g., NUBEAM), and empirical data where modeling is lacking (e.g., electron transport). An ensemble of models is generated to provide an indication of uncertainty. The optimization approach has been applied to propose optimal beam power, plasma shaping, and plasma current to achieve a target profile evolution. 

PP11.00165: Exploring higherorder effects in laserdriven ion acceleration via deep learning Blagoje Djordjevic, Andreas J Kemp, Joohwan Kim, Scott Wilks, Raspberry A Simpson, Ghassan Zeraouli, Elizabeth S Grace, Joshua Ludwig, Tammy Ma, Derek Mariscal Computer models of intense, laserdriven ion acceleration require expensive particleincell (PIC) simulations that may struggle to capture all the multiscale, multidimensional physics involved. We discuss an approach to ameliorate this deficiency, using a physicsinformed, multifidelity model that can incorporate physical trends and phenomena at different levels. As the base framework for this study, an ensemble of approximately 10,000 1D PIC simulations was generated to buttress a separate ensemble of hundreds of highfidelity, one and twodimensional simulations. Using transfer learning and multifidelity modeling in a deep neural network, one can reproduce the more complex physics at a much smaller cost. The networks trained in this fashion can in turn act as a surrogate model for the simulations themselves, allowing for quick and efficient exploration of the parameter space of interest. Standard figuresofmerit were used as benchmarks such as the hot electron temperature and peak ion energy, in addition to higherorder data such as the fields and particle phase space. These surrogate models are also useful for incorporating more complex scenarios, such as pulse shaping, that are challenging to model systematically let alone execute. We can rapidly identify and explore under what conditions dimensionality becomes a predominant effect as well as the transition between acceleration mechanisms. 

PP11.00166: Deep Learning with ICF Experimental Data Michael Pokornik, Andrew Maris, Shahab Khan, Luc Peterson, Kelli D Humbird With the recent recordbreaking yields from the latest shots at the National Ignition Facility (NIF), it is likely that we have not yet fully optimized the Inertial Confinement Fusion (ICF) experimental design. Machine learning has already demonstrated capability in ICF design discovery and is gaining traction in the community as a powerful tool. Artificial Neural Networks (ANN) have experienced success across numerous fields of science but often require large amounts of data to train and often overfit with small data sets. Due to the limited amount of ICF experiments, ANNs frequently train on data generated from numerical simulations. These simulations can be limited in capturing all the physics involved with experiments and may lose out on key features needed for performance prediction. 

PP11.00167: Multifidelity Regression of Sparse Plasma Transport Data Available in Disparate Physical Regimes Lucas J Stanek, Shaunak D Bopardikar, Michael S Murillo Physical data, such as material properties, are typically generated by several methods, including experiments and computations, in limited parameter regimes accessible to those methods. When datasets generated using such disparate methods are combined into one dataset, the resulting combined dataset is typically sparse with dense "islands" in a potentially highdimensional parameter space, and predictions must be interpolated among such islands. Using plasma transport data as our example, we use a nonlinear multifidelity Gaussianprocess regression framework that exploits physical data from multiple sources at multiple fidelities. We find that including data from multiple fidelities can reduce singlefidelityonly interpolation error by an order of magnitude resulting in predictions that are accurate across disparate regiems. We discuss how the multifidelity Gaussian process regression framework reveals the effectiveness of the lowfidelity data. When the lowfidelity model is insufficient, the framework reduces to singlefidelity Gaussian process regression. This framework suggests where data should be collected to generate more accurate predictions and can be readily applied to lookup tables for transport coefficients or equations of state that are often used in hydrodynamic simulations. 

PP11.00168: Explainable deep learning for the analysis of MHD activity in lockedmode disruptive pulses Diogo R Ferreira, Tiago A Martins, Paulo Rodrigues The locked mode amplitude is one of the most commonly used signals for disruption prediction in tokamaks. On a dataset from the JET baseline scenario, our results suggest that the simple application of a threshold on that signal yields a disruption predictor with more than 95% accuracy. It is wellknown that mode locking is one of the main disruption causes at JET; however, it is often too late to avoid a disruption by the time it is detected. In this work, we investigate the possibility of predicting the locked mode itself and, in particular, whether it is possible to apply machine learning to identify the MHD behavior that typically precedes mode locking. For this purpose, we trained a deep learning model over MHD spectrograms. The model is a Convolutional Neural Network (CNN) that receives a time window from the spectrogram and predicts whether the locked mode amplitude will exceed a given threshold. In addition, we use Class Activation Mapping (CAM) to explain why the model arrives at a certain prediction. The results suggest that the interruption of MHD activity followed by the resurgence of a mode at the q=2 surface are strong indicators that mode locking is about to occur, which is consistent with the literature. This work also suggests that neural networks can be useful, as interpretable machine learning models, to support the analysis of MHD activity. 

PP11.00169: Adaptive Generation of Machine Learning Data for Reduced FASTRAN Model Generation Mark R Cianciosa, Jin Myung Park, Wael Elwasif, Ross Whitfield Machine learning provides tools that can distill complex physics into a fast proxy function that still captures the underlying physics. A naive uniform sampling of the input parameter space can miss critical features of the output space. In addition, the full physics model can be prohibitively expensive to generate highresolution samples. By measuring changes in the model response with respect to the machine learning parameters, uncertainty and correlations of model parameters can found. This in turn is used to propagate uncertainty to the model predictions. In the regions of high uncertainty, the physics model can be resampled augmenting the initial data set. By continuously adapting the data and retraining, an optimal machine learning model is obtained through a smarter sampling of the input parameter space. This work presents an IPSbased workflow that couples machine learning model training with data generation. Using this workflow, neural network reduced models for tokamak system design studies are generated by adaptively sampling a theorybased IPSFASTRAN[1] integrated model. 

PP11.00170: Convolutional Neural Networks for MagLIF Stagnation Image Applications William E Lewis, Patrick F Knapp, Eric Harding, Kristian Beckwith Magnetized Liner Inertial Fusion (MagLIF) is a magnetoinertial fusion (MIF) concept being studied on the Zmachine at Sandia National Laboratories. MagLIF experiments employ imaging diagnostics that are used to constrain fuel plasma conditions and morphology. Analysis of these images involve tasks such as region of interest selection, background subtraction, and image registration of multiple images. While in principle these are not too difficult for an expert, manual treatment is tedious, increases risk of irreproducibility, and impedes quantification of uncertainty. To reduce the need for user input and time required to achieve common image analysis tasks, we present a convolutional neural network (CNN) based image segmentation able to detect pixels belonging to the stagnation column in Crystal Xray Imager (CXI) images. This enables more fully automated image analysis pipelines and collective assessments of many images, which may lead to physical insights. In particular, we utilize this CNN approach to demonstrate an automated background subtraction pipeline. This enables statistical analysis of both slowly varying background signal and random noise in a large ensemble of CXI images, which will lead to accurate models of background and noise that are commensurate with experiment. In addition, we conduct unsupervised clustering of background subtracted images using an image similarity metric. We look for physically meaningful clustering to assess the viability of chosen metric for imagetoimage comparison. 

PP11.00171: Data Representations for Virtual Diagnostics PeerTimo Bremer High fidelity computer simulations have long been used to help 
Follow Us 
Engage
Become an APS Member 
My APS
Renew Membership 
Information for 
About APSThe American Physical Society (APS) is a nonprofit membership organization working to advance the knowledge of physics. 
© 2024 American Physical Society
 All rights reserved  Terms of Use
 Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 207403844
(301) 2093200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 5914000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 200452001
(202) 6628700